Lenzinger Berichte 94/2018

Home , Carbon fibers, Defibrator, Organosolv

Scientific Reports on Wood, Pulp and Regenerated Cellulose

www.lenzing.com LENZINGER BERICHTE

Scientific Reports on Wood, Pulp and Regenerated Cellulose

Dedicated to Univ.-Prof. Dr. Paul Kosma on the Occasion of his 65th Birthday

„Älter werden ist wie einen Berg besteigen: Je höher man kommt, umso mühsamer wird der Weg, aber umso weiter wird der Blick.“

Ingmar Bergmann (1918 - 2007), schwedischer Drehbuchautor, Film- und Theaterregisseur

AU-ISSN 0024-0907 Published by Lenzing AG 4860 Lenzing, Austria Phone +43 7672 701-0 Editorial Board

Ao. Univ. Prof. Dr. Thomas Bechtold Universität Innsbruck, Research Institute of Textile Chemistry and Textile Physics, Austria

DDr. Haio Harms Kelheim Fibres GmbH, Germany (retired)

Univ. Prof. Dr. Paul Kosma University of Natural Resources and Applied Life Sciences, Department of Chemistry, Vienna, Austria

Univ. Prof. Dr. Herbert Sixta Aalto University, Department Chemical Pulping Technology, Espoo, Finland

Univ. Prof. Dr. Dr. Thomas Rosenau University of Natural Resources and Applied Life Sciences, Department of Chemistry, Vienna, Austria

The authors assume full responsibility for the contents.

Editor: Dr. Thomas Röder Printed in Austria, September 2018

Front page: Copyright Lenzing AG Fotos: Copyright Lenzing AG Layout: MMS Werbeagentur

Additional copies can be obtained by contacting the following address:

Dr. Thomas Röder c/o Lenzing AG, R&D, 4860 Lenzing, Austria Phone: +43 7672 701-3082, Fax: +43 7672 918-3082 E-mail: [email protected]

All papers are available online at: http://www.lenzing.com and http://www.lenzinger-berichte.com TABLE OF CONTENTS

I FOREWORD

SUSTAINABILITY AND BIOREFINERY

1 Basic Indicators for the Sustainability of European Forestry Peter Schwarzbauer and Fritz Wittmann 15 Development of an Organosolv Biorefinery Based on Nanolignin as Main Product Angela Miltner, Martin Miltner, Stefan Beisl, Walter Wukovits, Michael Harasek, Daniel Koch, Bettina Mihalyi, and Anton Friedl

ANALYTICS

27 FTIR Micro-Spectroscopy for Characterisation of Radial Substituent Diffusion in Cellulose II-Beads Danuta Mozdyniewicz, Andreas Schwaighofer, Karin Wieland, Martin Häubl, and Bernhard Lendl

CELLULOSE CHEMISTRY

33 Key Aspects of the Bleaching of Industrially Found Cellulosic Chromophores Nele Sophie Zwirchmayr, Ute Henniges, Thomas Röder, Antje Potthast, and Thomas Rosenau 43 Cellulose Carboxylate/Tosylate Mixed Esters: Dependence of their Physicochemical Properties on the Degree of Carboxylate Substitution Daniela C. Ferreira, Gustavo S. Bastos, Annett Pfeiffer, Denise F. S. Petri, Thomas Heinze, and Omar A. El Seoud

57 Kinetics of Cellulose Acylation with Carboxylic Anhydrides and N-acylimidazoles in Ionic Liquid/Molecular Solvent Mixtures: Relevance to the Synthesis of Mixed Cellulose Esters Thais A. Bioni, Naved I. Malek, and Omar A. El Seoud

FIBERS

67 Lignocellulosic Multicomponent Fibers Spun from Superbase-Based Ionic Liquids Michael Hummel, Yibo Ma, Anne Michud, Shirin Asaadi, Annariikka Roselli, Agnes Stepan, Sanna Hellstén, and Herbert Sixta

77 Man-Made Cellulosic Fibers via the Viscose Process – New Opportunities by Cellulose Carbamate R. Protz, G. Weidel, and A. Lehmann

85 Influence of CarbonizationAids on the Manufacture of Carbon Fibers based on Cellulosic Precursors Johanna M. Spörl, Frank Hermanutz, and Michael R. Buchmeiser TEXTILE APPLICATIONS

95 Continuous Alkali Treatments of Lyocell. Effects of Alkali Concentration, Temperature and Fabric Construction. Ján Široký, Avinash P. Manian, Barbora Široká, Richard S. Blackburn, and Thomas Bechtold 105 Chitin Coated Cellulosic Textiles as Natural Barrier Materials Auriane Freyburger, Yaqing Duan, Cordt Zollfrank, Thomas Röder, and Werner Kunz 115 Improvement of Dyeing Performance of Cellulose Fibers by Pre-treatment with Amino Cellulose Thomas Heinze, Thomas Wellhöfer, Kerstin Jedvert, Andreas Koschella, and Hendryk Würfel

RAW MATERIALS

123 Possibilities for Expansion of the Raw Materials Base in Lyocell Applications by Enzymatic Treatment Birgit Kosan and Frank Meister Foreword

The authors of the study “The New Plastics Economy: Rethinking the future of plastics“, which was presented at the World Economic Forum in Davos in 2016, rendered a great service to the society. The results of their work have gone around the world and made millions of people aware of environmental pollution caused by plastic waste. According to the study of the Ellen MacArthur Foundation, without action there will be more plastic than fish in the oceans by 2050.

Many of our major contemporary challenges – such as marine pollution – can only be solved with innovations, a core strength of the Lenzing Group. Regarding this problem, we are conscious of our corporate responsibility and are doing everything possible to make a positive contribution to mankind and to the environment by making sustainability a condition for every new development at Lenzing.

In the 2017 financial year, we invested more than EUR 55 mn in R&D, a top result in the sector both in absolute figures and in relation to revenue. With their outstanding performance, our approx. 190 employees working in process and product development not only ensure the economic success of the Lenzing Group, but they also make a significant contribution to solving many social and ecological problems.

The “Lenzinger Berichte” cover many important issues along the entire value chain, from wood to pulp up to the fiber and beyond. In this edition, you will find reports about sustainable forestry and the efficient use of resources, the further development of the biorefinery concept, the ecologically responsible production and further processing of wood-based fibers as well as new and innovative cellulose-based materials. In accordance of our new brand positioning “Innovative by nature”, I wish you sustainably inspiring reading.

Dr. Stefan Doboczky, MBA Chief Executive Officer Lenzing AG Lenzing, September 2018

Vorwort

Die Autoren der Studie „The New Plastics Economy: Rethinking the future of plastics“, die 2016 auf dem Weltwirtschaftsforum in Davos vorgestellt wurde, haben der Gesellschaft einen großen Dienst erwiesen. Die Ergebnisse ihrer Arbeit sind um die Welt gegangen und haben Millionen Menschen für die Verschmutzung der Natur durch Plastikmüll sensibilisiert. Wenn nicht gehan- delt wird, so die Studie der Ellen-MacArthur-Foundation, wird es in den Ozeanen 2050 mehr Plastik geben als Fische.

Viele der großen Herausforderungen unserer Zeit wie die der Meeresverschmutzung lassen sich nur mit Innovationen lösen, und diese sind eine zentrale Stärke der Lenzing Gruppe. Wir sind uns dabei unserer unternehmerischen Verantwortung bewusst und setzen alles daran, einen positiven Beitrag für Mensch und Umwelt zu leisten, indem wir Nachhaltigkeit zur Bedingung für jede neue Entwicklung bei Lenzing machen.

Im Geschäftsjahr 2017 haben wir mehr als EUR 55 Mio. in die Forschung & Entwicklung investiert – sowohl in absoluten Zah- len als auch in Relation zum Umsatz ein Spitzenwert in der Branche. Mit ihren hervorragenden Leistungen sichern unsere rund 190 Mitarbeiter/innen in der Prozess- und Produktentwicklung nicht nur den wirtschaftlichen Erfolg der Lenzing Gruppe. Sie leisten eben auch einen wesentlichen Beitrag zur Lösung vieler gesellschaftlicher und ökologischer Probleme.

Die Lenzinger Berichte decken wichtige Fragestellungen entlang der gesamten Wertschöpfungskette vom Holz über Faserzell- stoff bis zur Faser und darüber hinaus ab. In dieser Ausgabe finden Sie Berichte über nachhaltige Forstwirtschaft und die effizi- ente Nutzung von Ressourcen, die Weiterentwicklung des Bioraffinerie-Konzepts, die ökologisch verantwortungsbewusste Er- zeugung und Weiterverarbeitung von holzbasierten Fasern sowie innovative neue Materialien auf Cellulosebasis. Im Sinne unserer neuen Markenpositionierung „Innovative by nature“ wünsche ich Ihnen eine nachhaltig inspirierende Lektüre.

Dr. Stefan Doboczky, MBA Vorstandsvorsitzender Lenzing AG Lenzing, September 2018

I Congratulation to Univ.Prof. Dr. Paul Kosma

This fall, Univ.Prof. Dr. Paul Kosma from BOKU University Vienna, a longstanding cooperation partner of Lenzing AG, turns 65. His birthday is a welcome occasion for a short look back, sincere congratulation, and best wishes for the future.

Paul Kosma studied Technical Chemistry at Vienna University of Technology, followed by a PhD study in Organic Chemistry at the same institution, concluding both educational phases “with excellence”. In 1980 he joined BOKU University as “Universitätsassistent” at the Institute of Chemistry, where he obtained the habilitation (venia docendi) in organic chemistry in 1988. In 1983 and 1999 he stayed as a visiting scientist at the N.D. Zelinsky Institute in Moscow, Russia, and at the Department of Biochemistry, University of Edin- burgh, UK, respectively. In 1990, he was appointed full professor at BOKU University Vienna, in which function he stayed until today.

Between 1992-1998 and 2002-2013 Prof. Kosma acted as the head of the Department of Chemistry. He was scientific head of the ASPEX project (Antiviral Spot of Excellence Vienna, 2006-2009) and director of the “Christian Doppler Laboratory of Pulp Reactivity”, which was established and run in close cooperation with Lenzing AG. Since this time, long and fruitful cooperation has continued. This is also reflected by Prof. Kosma´s editorial board membership in “Lenzinger Berichte” – which makes this short retrospect the right place to say a big thank you also for that. Highlights of the “CD-lab” time were the „Japanese-European Workshop on Cellulose and Functional Polysaccharides“, Vienna 2005, and the „9th European Workshop on Lignocellulosics and Pulp”, Vienna 2006, both of which he organized and chaired.

With the establishment of the Christian Doppler Laboratory, he also laid the foundation for the institutionalization of cellulose and polysaccharide science at BOKU, which became one thriving branch of the “Chemistry of renewable resources” field, which he always supported, fostered and nurtured. In 2013, the Division of Chemistry of renewable resources” was founded, the destination of a long path that Prof. Kosma had opened and continuously paved. Also today, the Division of Organic Chemistry and its younger offspring, the Division of Chemistry of Renewable Resources remain connected by tight scientific links – but far more than that: by strong bonds of collegiality and friendship.

Prof. Kosma was BOKU delegate at FWF and head of the European Carbohydrate Society; he has been serving the scientific community as evaluator and peer for an incredibly high number of institutions, journals and research projects, and continues to do so today, also as the Austrian Delegate in the International Carbohydrate Organization (ICO) and European Carbohydrate Organization (ECO), Board Member of the Christian Doppler Research Society, and Member of the Austrian Agency for Scientific Integrity.

Paul Kosma is internationally recognized today as being one of the world-leading figures in carbohydrate chemistry, carbohydrate synthesis and glycochemistry, having published more than 270 scientific papers that were cited more than 5000 times.

The editorial team of “Lenzinger Berichte”, also on behalf of Lenzing AG as a whole and its many collaborators of Prof. Kosma's, would like to say “Happy birthday!” and “Thank you”.

On behalf of many congratulants, Dr. Thomas Röder, Dr. Axel Russler, Univ.-Prof. Dr. Dr. Thomas Rosenau

II In Memoriam Dr. Vasken Kabrelian

Cellulose scientist Dr. Vasken Kabrelian passed away at the 5th EPNOE International Polysaccharide Conference in Jena on September 23rd at the age of 57. He collapsed in front of the Lecture Hall of Jena’s Friedrich University, Germany and was dead at once – just one day before he wanted to give his talk on “Preparation of xylan derivatives from hemi-rich alkaline process lye”.

As a young Syrian scientist Kabrelian was part of the group of Prof. Werner Berger at Dresden University, Germany. There he earned a doctorate in organic chemistry entitled “Synthesis, characterization and experiments for applications of new dipolar O-basic organic solvents for cellulose”. He returned to Syria and finally worked as lecturer and researcher at Aleppo University. During a second stay in Germany in 2001 he assisted Prof. Dieter Klemm in his well-known cellulose research group at Jena’s Friedrich Schiller University.

After bearing up against the war in Aleppo for more than three years he decided to leave his home country together with his wife and their twins in 2014. His family belongs to the Armenian Christian minority and life had really become a daily struggle. They had to start all over again in Austria after their daring escape. He worked as a project manager at Kompetenzzentrum Holz GmbH for Lenzing AG, where he continued his scientific work. He is author of several patents and scientific publications.

Many appraised his cordiality. After all he still was a warm hearted, optimistic and pleasant person spreading happiness and always ready for a little joke. He saved his family and, thus, could not enjoy the fruits.

With Vasken Kabrelian we lost an experienced scientist, an esteemed colleague and a really good friend.

Team Lenzing R&D Team Kompetenzzentrum Holz GmbH

III LENZINGER BERICHTE 94 (2018) 01 – 13

Basic Indicators for the Sustainability of European Forestry

Peter Schwarzbauer* and Fritz Wittmann

University of Natural Resources and Life Sciences, Vienna. Institute of Marketing and Innovation, Feistmantel- straße 4, 1180 Wien, Austria, and Wood K plus Market Analysis and Innovation Research Team, Vienna, Austria * Contact: [email protected]

Abstract

This study focuses on one important set of indicators regarding the sustainable management of forests (SMF) – forest resource s. It is based on the latest internationally and partly nationally published forest inventory data. The developments of overall forest area, growing stock, increment, fellings and felling rates are covered for Europe as a whole (excl. Russian Federation), the EU28 and the total of European countries outside of the EU. In addition, forest resource changes were analyzed for single European countries (Austria, Germany, Czech Republic, Slovakia, Slovenia, Hungary and Belarus) as well as the Russian Federation, including – with limited data availability – also the specific resource developments for Norway spruce and beech. The general time-frame for the study was 1990-2015. Forest area and growing stock are increasing in all regions/countries. Fellings everywhere are below or far below net-annual increment; this underutilization is the main reason for growing stock increasing much more than forest area. In some countries, a shift from coniferous-resources to broadleaved-resources has become visible.

Keywords: European forest resources, sustainability indicators

Background and mandate

Sustainable management of forestry in Europe is guided that comparable data are available we aim for a longer- by stringent forest laws. However, outside of Europe term presentation of these developments over the last this is not well known. Therefore, the practice of sus- decades (whenever possible 1990-2015). tainable forest management (SFM) in Europe has to be documented and presented by available and tangible This report contains two different sets indicators. of data: One important set of indicators regarding the sustain- 1. The development of total forest resources (forest able management of forests contains forest resources. area, growing stock, increment, fellings, felling rate) The goal of this study is to present the development of by the following countries/country groupings: forest resources (area, growing stock, net-annual-incre- • Europe total (excl. Russian Federation) ment (NAI), fellings – including the ratio of increment • EU 28 and fellings [felling rate: fellings in % of NAI]) over • total of European countries outside the EU time in quantitative terms and – as far as possible given • single European countries: Austria, Germany, Czech consistent data availability – for the species Norway Republic, Slovakia, Slovenia, Hungary, Belarus spruce (Picea abies) and beech (Fagus sylvatica). Given • Russian Federation

1 LENZINGER BERICHTE 94 (2018) 01 – 13

2. The development of Norway spruce and beech re- Quantitative Basic Indicators sources (as far as consistent data are available) for for the Sustainability of European the countries: Forestry by Regions and Specific • Austria Countries • Germany • Czech Republic Europe total (excl. Russian Federation) • Slovakia From 1990-2015 forest resources in Europa have • Slovenia significantly grown; forest area by +9%, growing • Hungary stock by +40% and net-annual increment (NAI) by • Belarus +15%. The increase of growing stock, much above • Russian Federation the increase of forest area and increment is due to the underutilization of forests, which can also be seen in Results and Discussions the third part of figure 1: Fellings are much below NAI, the felling rates are 60-65%, which means that The latest international study with largely comparable only 2/3 of the sustainable potential is actually used. data on forest resources by European countries including the development over time and the main data source for this study is:

FAO/EFI (2015): State of Europe’s Forests 2015, published by the “Ministerial Conference on the Pro- tection of Forests in Europe – FOREST EUROPE Liaison Unit Madrid”. The data contained in this study are therefore not only comparable for all European countries as well as up-to-date but also official! The general time span covered is from 1990-2015, the last 25 years. However, there are some limitations. Not for each country all data are available. The regional totals therefore are always limited to the country data available. Another aspect is that the forest inventories in the various countries are not always conducted at the same time period(s). The country results were inter- and ex- trapolated in order to harmonize these data regarding the time context. Reported are the years: 1990, 2000, 2005, 2010, 2015. However, for some countries the data on resources are not covering the whole time frame (sometimes only 2000-2010). The specific data on Norway spruce and beech resources are not as consistent as the overall forest resource data. For Austria and Germany we rely on the published results of the national forest inventories (BFW for Austria, the Thünen Institute’s database for Germany) as well as on the country reports for the FAO Global Forest Resources Assessment 2015 (FRA). For the other countries we only rely on the respective country reports for the FRA [1]. The data contained in the FAO/EFI report as well as in the country reports for the FRA are transferred into Figure 1: Development of Forest Resources in Europe graphs, which form the main part of this study. total (excl. Russian Federation).

2 LENZINGER BERICHTE 94 (2018) 01 – 13

European Union (28) Total of European Countries other than From 1990-2015 forest resources in the European EU28 (excl. Russian Federation) Union have significantly grown; forest area by +9%, From 1990-2015 forest resources in Europa outside growing stock by +38% and net-annual increment of the EU have also grown significantly; forest area by (NAI) by +17%. The increase of growing stock, much +9%, growing stock by +47% and net-annual incre- above the increase of forest area and increment is ment (NAI) by +11%. The increase of growing stock, again due to the underutilization of forests, which can much above the increase of forest area and increment also be seen in the third part of figure 2: Fellings are is again due to the underutilization of forests, which much below NAI, the felling rates are 70-75%, which can also be seen in the third part of figure 3: Fellings means that only 3/4 of the sustainable potential is are much below NAI, the felling rates are only around actually used. 20%, and fellings have slightly declined over time.

Figure 2: Development of Forest Resources in the Figure 3: Development of Forest Resources in European European Union (28). countries other than EU28 (excl. Russian Federation).

[1] In case more detailed information on spruce and beech resources are necessary, the country experts and contributors to the FAO FRA 2015 need to be contacted: http://www.fao.org/forestry/fra/83059/en/

3 LENZINGER BERICHTE 94 (2018) 01 – 13

Austria Spruce Spruce still covers about 51% of the Austrian forest Total forest resources area available for wood supply, although the spruce From 1990-2015 forest resources in Austria have forest area declined between the 1990ies and 2007-09 significantly grown; forest area by +3%, growing stock by -8%; however, growing stock grew by +15%, net- by +25% and net-annual increment (NAI) by +6%. annual increment (NAI) by +16%. Although the felling The increase of growing stock, much above the in- rate of spruce increased from 70% to 88%, still less is crease of forest area and increment is again due to the harvested than annually grows. This is also the main underutilization of forests, which can also be seen reason that spruce growing stock increased, despite that in the third part of figure 4: Felling rates are spruce forest area declined. The decrease in spruce 60-90%, which means that not all of the sustainable forest area is not caused by overexploitation of spruce potential is actually used. resources, but results from increased regeneration of forests by more broadleaved species (see also chapter Beech).

Figure 4: Development of Forest Resources in Austria. Figure 5: Development of Spruce Forest Resources in Austria.

4 LENZINGER BERICHTE 94 (2018) 01 – 13

Beech Germany Beech covers about 10% of the Austrian forest area available for wood supply. All beech forest resource Total forest resources categories increased between the 90ies and 2007-09; From 1990-2015 forest resources in Germany have forest area by +8%, growing stock by +20%, net-annual grown, but at a low pace – increment has slightly de- increment (NAI) by +24%. The felling rate of beech creased; forest area increased by +1%, growing stock around 60% is much lower than for spruce, which is the by +34% and net-annual increment (NAI) slightly de- main reason that growing stock grew at a faster pace creased by - 0,1% between 2000 and 2010. The increase than forest area. The increase in beech forest area is of growing stock, much above the increase of forest mainly caused by increased regeneration of forests by area and increment is again due to the underutilization more deciduous species (see also chapter Spruce). of forests, which can also be seen in the third part of figure 7: Felling rates are 75-80%, which means that not all of the sustainable potential is actually used. The decrease of increment could be caused by overaged forests.

Figure 6: Development of Beech Forest Resources in Austria.

Figure 7: Development of Forest Resources in Germany

5 LENZINGER BERICHTE 94 (2018) 01 – 13

After the German reunion in 1990, the first forest in- Beech ventory including both (former) parts of Germany Beech covers about 15% of the German forest area took place in 2002, therefore, the availability of con- available for wood supply. All beech forest resource sistent data is mainly limited to the period 2002 to categories increased between 2002 and 2012; forest 2012 (most recent inventory). Growing stock data for area by +7%, growing stock by +10%. The felling rate spruce and beech are also available for 1990 (total of of beech around 70% is much lower than for spruce, West and East Germany). which is the main reason that growing stock grew at a faster pace than forest area. The increase in beech forest Spruce area is mainly caused by increased regeneration of Spruce still covers about 25% of the German forest forests by more deciduous species (see also chapter area, but spruce forest area has declined between 2002 Spruce). and 2012 by -8%; growing stock also declined in that period by -4%. Comparable data on NAI and fellings is only available for 2012 showing a felling rate of about 107%. However, the decline in spruce resources was more than compensated by the increase of broadleaved resources (see chapter Beech).

Figure 9: Development of Beech Forest Resources in Germany.

Figure 8: Development of Spruce Forest Resources in Germany.

6 LENZINGER BERICHTE 94 (2018) 01 – 13

Czech Republic Spruce In the Czech Republic about 61% of total growing Total forest resources stock and 75% of coniferous growing stock is spruce, From 1990-2015 forest resources in the Czech Re- which has increased from 1990 to 2010 by +15%. public have grown, but at a moderate pace; forest area by +1%, growing stock by +9% and net-annual increment (NAI) by +9%. Given the fact that felling rates have increased, but are still below 90%, the decrease in increment and growing stock since 2005 might be caused by overaged forests and/or a change in forest inventory methodology.

Figure 11: Development of Spruce Growing Stock in the Czech Republic (million m3).

Beech In the Czech Republic about 7% of total growing stock and 36% of broadleaved growing stock is beech, which has increased from 1990 to 2010 even by +45%.

Figure 12: Development of Beech Growing Stock in the Czech Republic (million m3).

Figure 10: Development of Forest Resources in the Czech Republic.

Regarding spruce and beech resources, consistent data is only available for growing stock.

7 LENZINGER BERICHTE 94 (2018) 01 – 13

Slovakia Regarding spruce and beech resources, consistent data is only available for growing stock. Total forest resources From 1990-2015 forest resources in Slovakia have significantly grown; forest area by +1%, growing stock Spruce by +21% and net-annual increment (NAI) by +33%. In Slovakia about 21% of total growing stock and Although growing stock slightly decreased since 2005, 70% of coniferous growing stock is spruce, which has the growth over the entire time period much above the increased from 1990 to 2010 by +21%. increase of forest area due to the underutilization of forests, which can also be seen in the third part of figure 13: Felling rates are 60-80%, which means that not all of the sustainable potential is actually used. At the same time and different from the countries/regions presented above increment itself increased more than area and growing stock: +33%!

Figure 14: Development of Spruce Growing Stock in Slovakia (million m3).

Beech In Slovakia about 32% of total growing stock and 59% of broadleaved growing stock is beech, which has increased from 1990 to 2010 even by +48%.

Figure 15: Development of Beech Growing Stock in Slovakia (million m3).

Figure 13: Development of Forest Resources in Slovakia.

8 LENZINGER BERICHTE 94 (2018) 01 – 13

Slovenia Regarding spruce and beech resources, consistent data is only available for growing stock. Total forest resources From 1990-2015 forest resources in Slovenia have significantly grown; forest area by +5%, growing stock Spruce by +54% and net-annual increment (NAI) by +53%. In Slovenia about 31% of total growing stock and The increase of growing stock, much above the in- 68% of coniferous growing stock is spruce, which has crease of forest area is due to the underutilization of increased from 1990 to 2015 by +49%. forests, which can also be seen in the third part of figure 16: The felling rates are below 40%, which means that about 60% of the sustainable potential is currently not used. At the same time and similar to the Czech Republic increment itself increased more than area and almost as much as growing stock: +53%!

Figure 17: Development of Spruce Growing Stock in Slovenia (million m3).

Beech In Slovenia about 31% of total growing stock and 68% of broadleaved growing stock is beech, which has increased from 1990 to 2015 by +57%.

Figure 18: Development of Beech Growing Stock

in Slovenia (million m3).

Figure 16: Development of Forest Resources in Slovenia.

9 LENZINGER BERICHTE 94 (2018) 01 – 13

Hungary Because of the minor importance of spruce in Hungary the FRA country report does not explicitly present data Total forest resources on this species. Regarding beech resources, the only From 1990-2015 forest resources in Hungary have available data is for growing stock. significantly grown; forest area by +15%, growing stock by +28%, and net-annual increment (NAI) has slightly decreased between 1990-2010 by -2%. The Beech increase of growing stock, above the increase of forest In Hungary about 11% of total growing stock and 13% area again is due to the underutilization of forests, of broadleaved growing stock is beech, which has in- which can also be seen in the third part of figure 19: creased from 1990 to 2010 by +7%. Fellings are below NAI, the felling rates are between 70 and 85%, which means that a part of the sustainable potential is not used.

Figure 20: Development of Beech Growing Stock in Hungary (million m3).

Figure 19: Development of Forest Resources in Hungary.

10 LENZINGER BERICHTE 94 (2018) 01 – 13

Belarus Because of the minor importance of beech in Belarus the FRA country report does not explicitly present data Total forest resources on this species. Regarding spruce resources, the only From 1990-2015 forest resources in Belarus have available data is for growing stock. significantly grown; forest area by +11%, growing stock by +59% and net-annual increment (NAI) by +29% from 1990 to 2010. No comparable data exist Spruce for fellings, but the fact that growing stock increased In Slovenia about 31% of total growing stock and so strongly shows that no overexploitation of forest 68% of coniferous growing stock is spruce, which has resources has taken place in Belarus (Figure 21). increased from 1990 to 2015 by +49%.

Figure 22: Development of Spruce Growing Stock in Belarus (million m3).

Figure 21: Development of Forest Resources in Belarus.

11 LENZINGER BERICHTE 94 (2018) 01 – 13

Russian Federation Regarding spruce and beech resources, consistent data is only available for growing stock. Total forest resources From 1990-2015 forest resources in the Russian Fede- ration have remained quite stable; forest area +/-0%, Spruce growing stock -1%, net-annual increment (NAI) +2%. In the Russian Federation about 13% of total growing Because fellings were reduced by 50% during this stock and 19% of coniferous growing stock is spruce, period, the felling rate also drastically declined from which has increased from 1990 to 2010 by +15%. 40% (1990) to 20% (2015) (Figure 23).

Figure 24: Development of Spruce Growing Stock in the Russian Federation (million m3).

Beech In the Russian Federation about 0,2% of total growing stock and 0,8% of broadleaved growing stock is beech, which has slightly decreased from 1990 to 2010 by -5%.

Figure 15: Development of Beech Growing Stock in the Russian Federation (million m3).

Figure 23: Development of Forest Resources in the Russian Federation.

12 LENZINGER BERICHTE 94 (2018) 01 – 13

Conclusions Slovakia N.N. (2014): Country Report Slovakia. Global Forest Sustainable forest management (SFM) has to keep the Resource Assessment 2015, Rome balance between three main pillars: ecological, economic and socio-cultural. SFM can be characterized Slovenia by a broad array of indicators, both quantitative and N.N. (2014): Country Report Slovenia. Global Forest qualitative. The focus here was on a set of quantitative Resource Assessment 2015, Rome indicators describing the development of forest resources: forest area, growing stock, increment, fellings Hungary and felling-rates. N.N. (2014): Country Report Hungary. Global Forest Other than global forest resources, the forest resources Resource Assessment 2015, Rome in Europe have significantly grown over the last 25 years. The increase of growing stock is much above the Belarus increase of forest area and increment due to the under- N.N. (2014): Country Report Belarus. Global Forest utilization of forests – fellings are and have been for a Resource Assessment 2015, Rome long time significantly below net-annual-increment. In Europe as a whole fellings amount to 60-65% of Russian Federation increment, which means that only 2/3 of the sustain- N.N. (2014): Country Report Russian Federation. able potential is actually used. Global Forest Resource Assessment 2015, Rome

Regarding this specific set of sustainability indicators, the results of this study show that (a) forest management in Europe can be labelled as „sustainable” and (b) that sustainable timber production could be even expanded.

Data Sources

General – European level FAO/EFI (2015): State of Europe’s Forests 2015. Minis- terial Conference on the Protection of Forests in Europe. FOREST EUROPE Liaison Unit Madrid, Madrid. http://www.foresteurope.org/docs/fullsoef2015.pdf (last accessed June 7th 2016)

Austria Bundesforschungs- und Ausbildungszentrum für Wald, Naturgefahren und Landschaft (BFW): Österreichische Waldinventur. http://bfw.ac.at/rz/wi.home (last accessed June 7th 2016) N.N. (2014): Country Report Austria. Global Forest Resource Assessment 2015, Rome

Germany N.N. (2014): Country Report Germany. Global Forest Resource Assessment 2015, Rome Thünen Institut: Dritte Bundeswaldinventur (BWI) 2012: Ergebnisdatenbank. https://bwi.info/ (last accessed June 7th 2016)

Czech Republic N.N. (2014): Country Report Czech Republic. Global Forest Resource Assessment 2015, Rome

13 LENZINGER BERICHTE 94 (2018) 15 – 26

Development of an Organosolv Biorefinery Based on Nanolignin as Main Product

Angela Miltner*, Martin Miltner, Stefan Beisl, Walter Wukovits, Michael Harasek, Daniel Koch, Bettina Mihalyi, and Anton Friedl

TU Wien, Institute of Chemical, Environmental and Bioscience Engineering, Getreidemarkt 9/166, A-1060 Vienna, Austria. * Contact: [email protected].

Abstract

Within the presented work the different aspects of the development of an organosolv biorefinery process with the focus on the production of sulphur free Nanolignin as main product are presented. Experimental investigations for the pre-treatment of the raw material using organosolv extraction as well as precipitation of lignin nanoparticles have been carried out. For the experiments wheat straw has been applied as the model substrate. Furthermore, design and analysis tools have been applied to investigate whole Biorefinery concepts using the in-house developed IDEAS-Engine. This toolbox has been applied for investigations on process design and optimisation, detailed optimisation of the extraction reactor, economics and life cycle assessments (LCA). The economic investigations have also been carried out for maize straw, soft and hard wood as raw materials. The optimized process conditions and equipment lead to Nanolignin with approximately 100 nm size and precipitation yields greater than 65 wt%. Energy integration using Pinch technology resulted in an exergy efficiency of 92%. Within the cost analyses Nanolignin showed the highest contribution to the revenues with a share of around 50% to 80% for hard and softwood, respectively, as well as 34-65% for wheat and maize straw. Within an economic risk analysis it was found that the risks of a negative profit for the Biore- finery concept is only 4-8% for the scenarios with soft wood and hard wood under the assumed conditions.

Keywords: Biorefinery, organosolv, Nanolignin, process development

Introduction

Lignocellulosic biomass is abundantly available and The integrated development of lignocellulosic Bio- offers a wide range of variability with many natural refinery concepts demands the consideration of various treasuries. Nevertheless these large fluctuations also aspects. General framework requirements for the pro- pose a challenge in regard to process development. cess development are appropriate process parameters Especially if a multi-feedstock and multi-product and conditions for the chosen raw materials leading to approach is chosen. The main constituents of ligno- a specific set of process streams. A development of appli- cellulosic materials are cellulose, hemicellulose, also cations for all process streams received during the pro- known as polyose, and lignin. However, also other cess is necessary with the aim of highest valuable associated components within the lignocellulosic material outputs. Diluted side streams that cannot be materials (e.g. extractives) are potential sources for value utilised economically for material production should be added products. Lignocellulosic biomass is tradition- used for energy production with the secondary objec- ally used as raw material in the pulp and paper industry tive of conserving nutrients which can be returned to which combines material and energetic use of the bio- the farmers as fertiliser for plant production whereby mass as well as solely energy production from ligno- closing the cycle. cellulosic biomass. But also other Biorefinery concepts When using an organosolv approach for the treatment can be based on lignocellulosic materials. of biomass, a wide variety of solvents like alcohols,

15 LENZINGER BERICHTE 94 (2018) 15 – 26

organic acids, phenol and cresol as well as esters, from the extract. In the “Precipitation” different preci- ketones and amines can be applied [1]. Many process pitation setups and conditions can be applied and variants have been investigated and developed to dif- nanoscaled lignin is produced. In the “Downstream ferent scales like ALLCELL, NAEM, ASAM and Processing”, resulting particles are separated from the ORGANOCELL processes [1, 2]. All of them have an suspension via centrifugation, dried and dispersed in integrated process development in common with the water. target of preserving the cellulose as main material product and extracting the other components in order to obtain additional products. In contrast, the approach of the presented research is to extract sulphur free lignin and process it to nanoparticles while utilising the nature of both the nanostructure and the chemical properties of native lignin from the production [3] to the application in various fields [4]. Along with lignin the cellulose stream as well as the mixed sugar stream is investigated to produce high value products. Side streams are dedicated for energy production. Based on long tradition and knowledge of the group, Figure 1: Schematic representation of the experimental process development and scale up of process ideas procedure from experimentally proven process steps on laboratory scale into pilot and semi-technical scale are done Organosolv pre-treatment with the help of various software tools. Process simu- The organosolv pre-treatment can be executed in dif- lation is used to derive substantial material balances. ferent apparatuses at the Research Division. For screen- Energy and exergy analyses are used for overall pro- ing experiments as well as kinetic experiments 45 ml cess optimization. Computational fluid dynamics stainless steel high pressure reactors (PARR, Model (CFD) is used for detailed optimization of the equip- 4716) are applied while detailed process investiga- ment design. Received process data are used for eco- tions are carried out in a 1 l stirred autoclave (Zirbus, nomic and ecological evaluation of the processes. HAD 9/16). In mid 2018 also an Extraction pilot plant This strategy leads to a toolbox called IDEAS-Engine will be available (Samtech Extraktionstechnik GmbH, equipped with design and analysis tools (e.g. process prototype) with a 10 l extractor that can be operated simulation, CFD, efficiency and cost analysis tools, up to 30 bar and 250 °C. LCA) applied for the support and speed up the pro- The optimum organosolv pre-treatment conditions are cess development. strongly dependent on the kind of biomass used and the desired main product of the pre-treatment step. In Materials and Methods the case of wheat straw and Nanolignin as desired product, experiments are typically carried out with Experiments 60 wt% aqueous ethanol mixture as solvent. The con- At present the organosolv Biorefinery experiments tent of wheat straw in the reactor is set between 8 wt% designed for Nanolignin production have all been and 20 wt% based on dry straw. The reactor is heated based on wheat straw as raw material. The wheat straw to temperatures between 120 °C to 220 °C within max. used was harvested in 2015 in Lower Austria, stored 45 min and held at this temperature for a defined time under dry conditions until use and was cut in a cutting span between 0 min and 60 min before cooling down mill, equipped with a 5 mm mesh, before pre-treatment. to room temperature. The solid and liquid fractions The composition of the straw was 16.1 wt% lignin and are subsequently separated using a hydraulic press 63.1 wt% carbohydrates consisting of Arabinose, (Hapa, HPH 2.5) at 200 bar and a centrifuge (Sorvall, Glucose, Mannose, Xylose and Galactose. Ultra-pure RC 6+) at 30,074 g for 20 min. The supernatant is water (18 MΩ/cm) and Ethanol (Merck, 96 vol%, un- analysed and stored at 5 °C. denatured) was used in the organosolv treatment and additionally sulfuric acid (Merck, 98 vol%) in the pre- Precipitation cipitation steps. The precipitation of the lignin dissolved in the organo- The experimental work is structured in three parts, solv extract is still under intense research. Applied illustrated in Figure 1. In the “Organosolv Pre-treat- precipitation setups combine the most commonly used ment” part, the lignin is extracted from the wheat straw methods solvent shifting and pH shifting and reduce applying an organosolv process and solids are removed the solubility of the lignin by decreasing the solvent

16 LENZINGER BERICHTE 94 (2018) 15 – 26

concentration and lowering the pH value [3]. The ex- available. Particle size measurements can be conduc- periments are conducted at room temperature. The ted in a Mastersizer 2000E (Malvern Instruments, volume ratio of extract to antisolvent is chosen in the UK) using laser diffraction in liquid suspensions range of 1:2 to 1:5, where the antisolvent consists of where two different measurements can be conducted: water set to a pH value of 2 to 7 with sulfuric acid. Up (1) Centrifuged particles dispersed in supernatant (with- to now three different setups have been investigated at out drying), (2) Centrifuged and dried particles redis- the Research Division: persed in water. a) “Batch” precipitation in a temperature controlled IDEAS-Engine stirred 250 ml beaker; 20 ml extract were placed in The design and development of concepts and plants in the beaker and 100 ml antisolvent were added with chemical engineering and related disciplines is per- a flowrate of 20 ml/min through an immersed anti- formed in strong interconnection between experimen- diffusion tip using a syringe pump. tal and simulative work. For a powerful and reliable b) “T-Fitting” precipitation consists of a T-Fitting and simulative approach, the research division Thermal a 20.4 cm long pipe with inner diameter of 3.7 mm Process Engineering and Simulation fully relies on a and a metering valve at its end. Extract and antisolvent toolbox called the IDEAS-Engine which is an acronym are pumped into the T-Fitting via two syringe pumps standing for Integrated Design and Analysis System. with a flowrate of 4 ml/min (20 ml total volume) and The exhaustive simulation of every aspect of a novel 20 ml/min (100 ml total volume), respectively. The process concept or an innovative plant layout necessi- precipitate is collected in a 250 ml beaker without tates the use of very diverse modelling and simulation stirring. The whole setup is temperature controlled. tools with a high degree of integration and interaction. c) “Static Mixer” precipitation uses the setup (b) but Thus, the IDEAS-Engine covers a multitude of rele- the pipe after the T-Fitting is equipped with static vant aspects like material property data, process simu- mixing elements (Striko, Helical DN4) that redirect lation, process analysis, CFD simulation and cost ana- the flow 30 times. Total volume and flowrates were lysis (Figure 2) which are usually executed in a identical to setup (b). repetitive and iterative approach. In every aspect, the development and validation of new models and algo- More details on the setups can be found in [5]. rithms as well as the gaining of expertise in their application are major tasks. Downstream processing The suspension gained in the precipitation step containing the Nanolignin particles is centrifuged immediately after precipitation to separate the lignin particles. For particle separation a Thermo Scientific, Sorvall RC 6+ centrifuge (typically used with 30,074 g), a Sigma 4k15 centrifuge (typically used with 24,104 g) and a Thermo WX Ultra 80 ultracentrifuge (typically used with 288,000 g) are available. The supernatant is decanted, collected and analysed.

Analytics The organosolv extract is typically analysed for carbohydrates, lignin and degradation products. Details can be found in [5] and [6]. Dried precipitate is investigated in terms of molecular mass via High Performance Size Figure 2: The IDEAS-Engine for design of innovative pro- Exclusion Chromatography (HPSEC) using an Agilent cesses and plants in chemical engineering and related disci- 1200 HPLC system equipped with a UV-Detector. Its plines chemical composition is analysed via ATR-FTIR absorption obtained with a Bruker Vertex 70 Fourier The assessment of material property data ranges from Spectrometer. The particle size and shape of the lignin simple data extraction from compendia or trustworthy particles are investigated in a scanning electron micro- web resources like NIST Chemistry WebBook [7] over scope (SEM) (Fei, Quanta 200 FEGSEM). For the in- well-accepted regression models for engineering pur- vestigation of the ζ-Potential as well as the particle poses and ends with highly predictive methods like size a ZetaPALS (Brookhaven Instruments, USA) is molecular dynamics simulation and ab-initio quantum

17 LENZINGER BERICHTE 94 (2018) 15 – 26

chemical simulation. Process simulation is performed Process Simulation with several different commercial simulation environ- Process simulation is employed for process design and ments like Aspen Plus, gPROMS and IPSEpro. The scale up to a plant capacity that could operate in the outcome of this step are self-consistent mass- and area of Lower Austria. The mass and energy balances energy balances of the process at different levels of de- obtained from process simulation allowed further in- tail. These balances form the input for the subsequent vestigation of the organosolv process via Pinch and process analysis step like heat and energy minimization, exergy analysis. design and optimisation of heat exchanger networks, A simulation model for the organosolv process was the analysis of exergy flow and Life Cycle Assess- developed in the simulation software Aspen Plus ment (LCA). (V7.3.2, Aspen Technology Inc., 2012), implement- A very powerful tool for the development of processes, ing experimental results and complemented with litera- plants and apparatus is Computational Fluid Dynamics ture data to balance all involved process steps. (CFD). CFD can be used as a versatile tool e.g. begin- NRTL-RK (Non-Random Two-Liquids Redlich-Kwong) ning from early assessment of chemical reactor de- property method is used to calculate phase equilibrium velopment and ending with the final optimisation of a and thermodynamic properties. Based on a plant capa- tube bend against abrasion and blocking. The IDEAS- city of 100 kt/year and assuming 8000 h/year operation Engine includes different commercial and open-source time, a mass flow of 12,5 t/h wheat straw is mixed with codes for pre-processing, simulation and post-proces- organic solvent (ethanol/water: 60 wt%/40 wt%) and fed sing. Last but not least, plant investment and opera- to the REACTOR (Figure 3). After the pre-treatment, the tional costs, profitability, return on investment, net REACTOR content is discharged and sent to the solid/ present values and similar economic key performance liquid separation unit (FILTER). Solid fraction is sepa- indicators are calculated in a concise cost analysis rated from liquid with a dry matter (DM) content of step based on standards like VDI 2067. Thus, the 30 wt% and then washed to prevent lignin re-precipita- IDEAS-Engine provides the tools and workflow for tion as well as ethanol loss. Wheat straw liquid phase an all-integrating, consistent and comprehensive de- is used for lignin recovery and isolation. Dissolution velopment of novel process concepts and plants. reactions and process conditions are reported in [8].

Figure 3: Organosolv process flowsheet. Base case without membrane separation: red lines, Process case with membrane separation: blue lines [8]

The liquid stream is treated in two different scenarios: [11], was used. The Mathematica (V6.0.3.0, Wolfram • direct precipitation (Figure 3, red path) Research, Inc., Champaign, USA, 2008) based software • pre-concentration prior lignin recovery (Figure 3, tool consists of sub-programs, created to perform: blue path) applying a nanofiltration membrane (NF • pinch analysis MEM) [9] • synthesis of heat exchanger networks Residual liquid is sent for solvent recovery via distil- • retrofitting of existing heat exchanger networks. lation column [10]. Additionally to Pinch analysis and heat integration, For Pinch analysis and heat exchanger network (HEN) exergy analysis is applied to further understand the synthesis, software TVTHENS developed at TU Wien process efficiency [12]. Contrary to the energy, exergy

18 LENZINGER BERICHTE 94 (2018) 15 – 26

can be lost due to process irreversibilities. Therefore, Fluid flow through the meshed packed bed was simu- exergy analysis gives information of kind and quantity lated using the open-source CFD package Open- of heat and material losses in the process as well as to FOAM®. A steady state incompressible solver was justify the products quality. selected (simpleFoam) based on the semi-implicit method for pressure linked equations. Residence time CFD Simulation distributions were calculated using scalarTransportFoam The process efficiency of unit operations along ligno- which solves a scalar transport equation based on the cellulosic biorefinery concepts is governed strongly by velocity field derived from simpleFoam. The liquid equipment design. Process intensification to improve used in the simulations was a 60% volume mixture of phase contact, to reduce solvent consumption and to ethanol/water entering the bed from bottom (z direc- maximize lignin extraction with minimum by-product tion) at 0.01 m/s and exiting from top. In this case, the generation is an important design goal with strong im- velocity regime was laminar and the wall-treatment pact on the selection of downstream unit operations for was no-slip; as a start, pressures and temperatures high product purity. The extractive separation of lignin were set to ambient, and the viscosity for the mixture from cellulose and hemicellulose is often carried out in was adapted accordingly. packed beds, where a uniform distribution of the sol- As shown previously [14], CFD – in combination with vent in the packed bed helps to maximize the contact laser-based flow measurement and Raman spectro- time between solvent and biomass particles or straw scopy – can also be applied to investigate mixing fibres and increase the efficiency. Usually, the solvent processes. With regard to organosolv biorefinery, this mixture is passed over the biomass packed bed. CFD approach could be used to get more insight into the simulations can give a very detailed view about the precipitation of Nanolignin. flow behaviour, uniformity and distribution in these packed bed systems. Process Analysis The analysis of random packings of particles or fibres via Life Cycle Assessment poses a great challenge on CFD as the creation of the Along with the presented design and optimisation of complex, unstructured geometry is usually not pos- the process using simulation tools, also the environ- sible applying standard CAD tools. A code based on mental impact has to be analysed to obtain a sustain- discrete element modelling (DEM) developed at TU able development. Life Cycle Assessment (LCA) is Wien using particle clumps to consider non-spherical carried out for different process concepts of the inno- or fibrous biomass particles was used in combination vative organosolv Biorefinery with Nanolignin as a with the open-source CFD tool OpenFOAM® for the main product, in order to investigate and minimize its simulation of the flow [13]. The filling of the extractor environmental implications. The main objective of this geometry with arbitrary non-spherical finite geometries methodical assessment is to provide information on the representing the biomass pieces was performed by an possible environmental effects connected to alternative automated procedure which includes STL file creation process concepts and product qualities within the Bio- (representing the geometry) and the computation refinery. Based on the gained impact informations well- of the mesh for the computational domain through founded decisions can be made for the further develop- SnappyHexMesh. ment of the technology. An example of an extractor geometry filled with fibres The LCA method requires life cycle thinking, including (two different sizes) is shown in Figure 4. raw material production, upstream and downstream process considerations as well as possible applications, use and end of life scenarios. A hotspot analysis in different impact categories highlights the critical points from an environmental perspective and enables a more holistic process development. The LCA method is standardized by ISO14040 and ISO14044 and consists of four consecutive stages (see Figure 5) with iterative nature [15, 16]. Goal and scope definition is followed by gathering of the Life Cycle Inventory, leading to the Life Cycle Impact Assessment which is followed by the interpretation phase. Findings in every stage can trigger the need to return to other Figure 4: DEM created packing using particle clumps stages and adapt the chosen scope, goal and methodo- (a, left) – STL file generated from DEM filling (b, right) logical decisions.

19 LENZINGER BERICHTE 94 (2018) 15 – 26

from the economic calculations are used to compare the costs of Biorefinery products with prizes of commercial products. Beside this well-known field of application for economic calculations and cost analysis, it is also important to consider economic aspects at an early stage of process development in order to support the identification of promising processes and product scenarios that are still under development. A model has been developed that allows the comparison of different innovative Biorefinery scenarios [17]. Within the model key performance indicators (KPI) for expenses and revenues are calculated and lead to a KPI representing the profit of the scenario. To identify the expenses KPI, the model of “levelized costs of electricity”, known from the energy production sector [18], that offers the possibility to compare long term investigations for systems with different cost structures, has been adapted. Literature data and marked data for investment costs, material costs, personnel costs and other costs have been used together with technical know-how to calibrate the model for the different Biorefinery scenarios. The gained expenses KPI also considers uncertainties as it bases on the weighted average cost of capital (WACC). Figure 5: Stages of a life cycle assessment, adapted from The revenue KPI is also based on literature data for ISO14040 [15] revenues for each product or for comparable products if the revenue for the intended innovative product is not For the life cycle inventory, laboratory experiments and yet known. In the latter case it is recommended to use a simulation results at first hand, are combined to gather span for the revenue rather than a single value. This is the best possible quality of inventory data at this stage especially true in the case of Nanolignin as a product of the technology development. For modelling life cyc- as it is not yet known, which final product will be gained le inventory the LCA software GaBi is applied. Data- from Nanolignin. Thus, the expectable revenues are also sets from commercial databases developed by the com- unknown at the moment. Revenues for Nanolignin pany thinkstep (www.gabi-software.com) are used for were assumed with values starting from 3.300 EUR/t. background data requirements. The other products of the Biorefinery concepts (cellulose, hemicellulose and normal sized lignin) have also Cost Analysis been assigned with revenue ranges [17]. Costs an economics always play an important role in Furthermore, the risks that are connected with inno- industry. The best engineered and most sustainable Bio- vative developments are considered in the model. refinery concept will not be implemented if it does not This is done by applying distributions (e.g. normal or proof to be economic. Therefore all Biorefinery con- triangular distribution) to the defined ranges for costs cepts developed within the Research Division are also and revenues and by carrying out Monte Carlo Simu- investigated from an economic point of view. We can lations. The results from these simulations lead to carry out these investigations using different tools probabilities for positive and negative profit KPIs that depending on the problem to be solved. Economic can be used as basis for decisions. calculations can be implemented in MS Excel® based on standards like VDI 2067. But also the software tools IPSEpro PSEconomy and Aspen Process Economic AnalyzerTM are used to carry out profitability calculations. These calculations are applied to identify optimization potentials used as input for experimental optimization and process simulation. They also form the basis to identify the most economic Biorefinery concept within some variants. Moreover the results

20 LENZINGER BERICHTE 94 (2018) 15 – 26

Results and Discussion

Experiments Preliminary experiments investigated the influence of the ethanol content during organosolv pre-treatment of wheat straw on the extraction yields of lignin and sugars [19]. The results showed the best lignin yields at ethanol concentrations of 50-60 wt%. In contrast, the sugar yields are maximized when the extraction is carried out with pure hot water and decreases considerably with increasing ethanol content. Thus, it is evident that the pre-treatment step has to be optimized for the desired biorefinery concept. If both Figure 6: Example of a scanning electron microscopy (SEM) sugars and lignin yields are claimed to be maximized, a image of dried lignin particles staged pre-treatment combining a pure water extraction step with an ethanol organosolv extraction step should solvent leads to maximum yields around 50 wt% while be developed. pH 2 antisolvent leads to yields greater than 65 wt%. Recent experiments have focused on the production Determination of the molecular mass of the Nanolignin of lignin particles in micro- to nanoscale. The pre- particles showed that the precipitation setup has only treatment process has, therefore, been implemented as minor influence on the molecular mass of the Nano- organosolv process with an aqueous ethanol solution lignin. The weight average molecular weight for the with 60 wt% ethanol. The experiments have been car- three investigated precipitation setups lies in the range ried out in the 1 l stirred autoclave with 8.3 wt% dry of 1.600 to 1.800 Da [5]. wheat straw content in the reactor and at a temperature of 180 °C that was held constant for 15 min. Details on Process Simulation the experiments are given in [5]. The resulting organo- Simulation is performed for a constant feedstock flow solv extract contains 0.677 g/l total carbohydrates and rate of 12.5 t/h wheat straw. Besides the base case (no 6.62 g/l total lignin. pre-concentration via membrane unit) a stage cut of This organosolv extract was used as feed for the three 15%, 50% and 80% is considered for the membrane precipitation setups described above. The precipita- unit. For ethanol and water, no rejections were observed tion experiments were carried out at 25 °C and with a experimentally, so the split factor was equal to the constant volume ratio of extract to antisolvent of 1:5. stage cut. For lignin, ash, and proteins, split factors of The antisolvent consisted of water set to a pH of 2 or 1%, 10%, and 50% of the stage cut were used. For the 5 by adding sulfuric acid. main sugars xylose, arabinose and glucose, the used It could be shown that the precipitation with the “Static split factors were 77.3%, 80.6% and 52.6% of the Mixer“ setup resulted in the smallest particle of 386 to stage cut. Based on the assumed feedstock input it 463 nm (for undried and dried particles, respectively, was possible to obtain around 26.2 t/h wet cellulose/ precipitated with an antisolvent with pH 2) followed hemicellulose stream as well as 2.5 t/h of wet lignin. by the “T-Fitting” setup and the “Batch” setup [5]. Both output streams are containing about 70% water. The width of the particle distributions from these Following Table 1 a pre-concentration step via nano- measurements on the other hand is smallest for the filtration (NF) brought some savings in the fresh water “Batch” setup followed by the “T-Fitting” setup and and acidified water demand. Furthermore, under the the “Static Mixer” setup. considered process settings pre-concentration prevents The experiments with the “Static Mixer” setup have ethanol losses. However, detailed experimental inves- also been carried out with an antisolvent with pH 5. tigations are necessary to confirm findings and lignin The comparison of the results from both “Static Mixer” quality [9]. experiments showed that the antisolvent with pH 5 leads to a significant reduction of the median diameter of the particles to approx. 100 nm [5]. These particle size results measured with the Mastersizer 2000E are supported by SEM images (Figure 6). The precipitation yields (ratio of mass of precipitate after drying to the lignin amount in the organosolv extract) showed that the precipitation with pH 5 anti-

21 LENZINGER BERICHTE 94 (2018) 15 – 26

Table 1: Basic mass balance for all process versions based on 12.5 t/h straw input. Values given in t/h

Basic energy balances are calculated for original (non- integrated) process designs, including thermal energy for fractionation and solvent recovery, not considering the drying of the final products. Obtained results are in a good agreement with studies reported in the literature (similar process designs, but less process steps considered), determining a specific energy demand for heating of 11.84 MJ/kg dry biomass [20] and 11.6 MJ/kg dry biomass [21]. While base case (V1) and process option V2_15 obtain a considerably higher specific heat demand than reported in literature, cases V2_50 and V2_80 are comparable or lower than reference processes (13.45 and 9.45 MJ/kg dry biomass, respectively; non-integrated) [8]. At this point, it is shown that utilization of NF could bring some advantages in energy savings without implementation of any heat integration. Further investigations show that Pinch analysis and heat integration has a discernible impact on improve- Figure 7: HEN for different process versions obtained from ment of process energy efficiency. Figure 7 illustrates TVTHENS the optimum HEN design suggested by the TVTHENS software. Version V1 shows lowest capability for heat integration, with minimum number of exchanging units and inability to integrate the column reboiler. Conversely, in V2_80, nanofiltration and its high performance could give a possibility to fully integrate the column reboiler and provide necessary heat from the process streams via HEN, resulting in a specific energy demand of 12.54, 8.16, 6.37 and 2.88 MJ/kg dry biomass for process options V1, V2_15, V2_50 and V2_80, respectively [8]. Figure 8: Exergy efficiency=Exergy_out/Exergy_in

22 LENZINGER BERICHTE 94 (2018) 15 – 26

Calculated exergy efficiency is presented in Figure 8. V1 has the lowest simple exergy efficiency of 76% due to exergy destruction, most probably in the solvent recovery block. That further leads to a low energy efficiency of the process, since thermal energy available in the process cannot be efficiently exchanged between process streams. Process version V2_80 ex- presses the highest exergy efficiency due to the minimization of energy consumption in the solvent recovery block. Only 8% of exergy is destroyed in this process version. Exergy loss is a consequence of non- Figure 9: Axial and radial porosity profiles plotted from the well designed process units (e.g. reactor, washing bottom to top and from the centre to the wall, respectively unit, separation etc.). However, at this point, it was not possible to locate these irregularities. For this The velocity field was then calculated using anin- purpose, exergy analysis of each process step has to compressible flow solver of OpenFOAM® at an aver- be performed [8]. age inlet velocity of 0.01 m/s. In Figure 10 velocity contour plots at different heights from bottom of the CFD Simulation packed bed are shown. Some high velocity spots can As an example to demonstrate the scope of CFD re- be observed close to the walls which show the chan- lated methods to investigate organosolv extractors, a nelling in the bed – the effect is more pronounced at straw fibre filled fixed bed extractor vessel was cho- the top of the extractor where the fibre packing is less sen. The investigated cylindrical extractor geometry compact. Channelling may result in short contact time was approximately 150 mm in diameter and around between straw fibres and liquid and consequently 100 mm in height. It was filled with 50% fibres of lower extraction performance. 30 mm length and 5 mm diameter and 50% fibres of The computation and analysis of the residence time 40 mm length and 4 mm diameter. The porosity pro- distribution (RTD) is a proper means to further in- files are a perfect indicator to analyse significant vestigate channelling and by-passing effects. By the bypassing or channelling. Figure 9 shows the average application of a step function concentration pulse axial porosity and radial porosity of the bed. The over- at the inlet the average concentration at the outlet is all average porosity was around 0.61 showing a observed with time. Wide RTDs are a good indica- much lower packing density compared to more sphe- tion for channelling and/or by-passing. For three rical particles. The porosity significantly increases different inlet velocities RTD curves were extracted towards the wall and towards the top of the extractor from the flow fields. As shown in Figure 11 (next vessel. This indicates the risk of wall channelling, page), straw fibre packings obviously show a wide however, with low flow velocities typically applied in RTD which can be narrowed through higher flow ra- fixed bed extractors, the channelling effect may be tes through the extractor. Further results of this study usually less significant. Also, the packing model does will be shown in [22]. not account for fibre softening and re-arrange- ment effects during a typical organosolv extraction process yet.

Figure 10: Velocity magnitude contours at the bottom, the centre and the top of the extractor

23 LENZINGER BERICHTE 94 (2018) 15 – 26

• organosolv extraction of either hard wood, soft wood, wheat straw or maize straw • organosolv extraction is carried out with an aqueous ethanol solution with 60 wt% ethanol • 99% of the ethanol can be recovered within an optimized recovery plant like the process presented in [8] • products gained in the organosolv extraction are (with varying contents depending on the used raw material): Figure 11: Residence time distribution (RTD) for three dif- - a lignin fraction, ferent inlet velocities - a hemicellulose fraction, - a cellulose fraction containing some lignin, hemi- Process Analysis cellulose and other constituents and via Life Cycle Assessment - a residual fraction First results are presented from the methodological • 50% of the lignin fraction can be processed to lignin perspective of LCA application in the process de- particles in nano scale velopment. • costs are considered for: Supporting the development of the Biorefinery by - investments making the environmental impacts available was found - raw material, ethanol, water to be the most suitable goal of the assessment. The - heat flexible nature of the LCA method matches the need - personnel to adapt to the progress within the process develop- - other costs ment of the Biorefinery. If new insights are gained and varied within defined ranges from experimental setups or simulation results, they • revenues are considered for: can be incorporated quickly to the developed life cycle - lignin as nanoscale particles as high prize product model of the Biorefinery. - lignin in non-nanoscale form used as combustion Burden allocation is an issue at several stages of the fuel Biorefinery. In case more than one product is gained, - Hemicellulose fraction used for heat production the question of how to distribute the environmental - mixed cellulose fraction used for heat production impacts between these products has to be answered. (worst case) or as input material for pulp and paper The wheat straw feedstock is an agrarian byproduct industry (best case) and the environmental impacts from agricultural pro- and varied within defined ranges cesses can be accounted between corn and straw with • the WACC is analysed with 0%, 6% and 12% different approaches. Furthermore the intermediate side products of the Biorefinery, cellulose and hemi- The analysis of the costs KPI showed that the invest- cellulose, have to be handled in a fair way. ment costs have a main influence on the levelized All emissions are related to the functional unit, which costs of production in all scenarios. In the straw based is defined in the goal and scope stage. For thede- scenarios heat and personnel are the major cost factors velopment phase of the Biorefinery it was considered within the costs of operation. In the scenarios based as sufficient to use a fixed physical parameter of the on hard and soft wood the raw material costs are the yielded product Nanolignin as the functional unit. For main costs of operation followed by heat and per- further assessments of comparative investigations sonnel. The total costs for the straw based concepts with other products, a functional unit has to be defined are significantly lower than the costs for hard wood that enables a meaningful comparison. and soft wood scenarios. This means that from the costs point of view the straw based scenarios are the Cost Analysis most promising ones. The cost analysis to support strategic decisions for the The analysis of the revenues for hard wood and soft development of a biorefinery process has been carried wood showed that the revenues for Nanolignin as the out for the following concepts (details can be found highest value added product in the concepts have the in [17]: major share of the total revenues ranging from around 50% to 80% depending on the wood and the revenue scenario. The second largest share has the cellulose fraction with around 15-45%.

24 LENZINGER BERICHTE 94 (2018) 15 – 26

For the straw based scenarios the results are more the handling of a multitude of relevant aspects like diverse. Wheat straw shows a share of Nanolignin on material property data, process simulation and pro- the revenues of around 44-65% and maize straw a share cess analysis. Process simulation has been applied of approx. 34-61%. This lower share is caused by the for investigating the recovery of the used solvents. much lower ratio of lignin content to cellulose content A nanofiltration process has been simulated as con- in the raw materials compared to hard wood and soft centration step for a solvent recycle stream with wood. Consequently, the share of the cellulose frac- recycle rates of 15, 50 and 80% of the solvent. This tion on the revenues is higher with around 26-53% for recycle stream is led directly back to the extractor wheat straw and around 34-61% for maize straw. without the need of a distillation step. This measure Both results can be combined to the profit KPI. This leads to a reduction of energy demand as well as a analysis shows that the straw based scenarios have a simplified design of the needed distillation column. negative profit for most of the investigated cases. The CFD simulation has been applied for detailed design profit for soft wood is slightly higher than the profit and analysis of the core unit within the Biorefinery for hard wood. Moreover, the Monte Carlo simula- concept, the extraction vessel. The liquid flow field tions showed that the risk for a negative profit is could be optimised and along with that the maximum around 50% for the wheat straw based scenarios. The utilisation of the extractor volume was ensured. maize straw based scenarios have a risk of a negative With the application of the Pinch technology on the profit of more than 70%. In contrast, the risks ofa Biorefinery concept, an energy optimization has been negative profit are around 4-6% for soft wood and achieved. On this basis the exergy analyses reveal an around 5-8% for hard wood. exergy efficiency of 92% or 8% loss of exergy. Based on the results it can be concluded that the in- The cost analysis for Biorefinery concepts with wheat vestigated scenarios with wheat and maize straw as raw straw and maize straw as substrates showed a share of material are not profitable under the maid assump- Nanolignin on the revenues of 44-65% and 34-61%, tions. It would, therefore, be necessary to identify respectively. The analysis for hard wood and soft higher valued applications for the hemicellulose and wood showed that the revenues for Nanolignin have a cellulose fraction in order to obtain higher revenues share on the total revenues ranging from around 50% and gain positive profits. This could, for example, be to 80%. The second largest share on the revenues has achieved by utilizing the hemicellulose fraction as the cellulose fraction with around 15-45%. A risk input for chemical or biochemical synthesis or by analysis identified that the straw based scenarios have improving the properties of the cellulose fraction by a high risk of a negative profit and need adaptions of optimized pre-treatment conditions. Additionally, a the assumed utilisation scenarios for the cellulose and reduction of the heat demand by heat integration or a especially for the hemicellulose fraction. In contrast, coupling with other industrial plants making use of the risks of a negative profit for the soft wood and their waste heat could increase the profitability. hard wood scenarios are around 4-6% and around 5-8%, respectively, and show high potential under the Conclusions made assumptions. The LCA calculations for the developed Biorefinery For the valorisation of the main constituents of ligno- process are still under investigation but first methodo- cellulosic materials cellulose, hemicellulose and lignin logical aspects have been presented. an organosolv Biorefinery approach using an ethanol/ water-mixture has been investigated. As a model sub- Acknowledgement strate wheat straw has been used as raw material. The leading product in the developed concept are sulphur The authors acknowledge the TU Wien University free native lignin micro- and nanoparticles. Library for financial support through its Open Access In extensive series of experiments in 45 ml and 1 liter Funding Program. Scanning electron microscopy high pressure reactors the organosolv pre-treatment image acquisition of the lignin particles was carried conditions have been optimised and will be used in a out using facilities at the University Service Centre 10 liter extractor for the production of enough sample for Transmission Electron Microscopy, TU Wien, material for further application tests. Due to optimized Austria. process conditions and equipment optimization for the precipitation of Lignin particles, yields greater than 65 wt% are achieved. The whole process development is supported by the in-house toolbox IDEAS-Engine which allows for

25 LENZINGER BERICHTE 94 (2018) 15 – 26

References [12] Modarresi, A., Kravanja, P. Friedl, A., Pinch and exergy analysis of lignocellulosic ethanol, bio- [1] Muurinen E., organosolv pulping – A review and methane, heat and power production from straw, distillation study related to peroxyacid pulping. Applied Thermal Engineering, 2012. 43: p. 20-28. PhD Thesis 2000, University of Oulu. [13] Haddadi, B., Jordan, C., Norouzi, H.R., Harasek [2] Yawalata D., Catalytic selectivity in alcohol pulp- M., Investigation of the pressure drop of random ing of spruce wood. PhD Thesis 2001, Universi- packed bed adsorbers, Chemical Engineering ty of British Columbia. Transactions, 2016. 52, 439-444. [3] Beisl, S., Miltner, A., Friedl, A., Lignin from [14] Haddadi, B., Gasser, C., Jordan, C., Harasek, M., Micro- to Nanosize: Production Methods. Int. J. Lendl, B., Simultaneous Laser Doppler Veloci- Mol. Sci. 2017. 18(6), 1244. metry and stand-off Raman spectroscopy as a [4] Beisl, S., Friedl, A., Miltner, A., Lignin from novel tool to assess flow characteristics of pro- Micro- to Nanosize: Applications. Int. J. Mol. cess streams. Chemical Engineering Journal, Sci. 2017. 18(11), 2367. 2018. 334, 123-133. [5] Beisl, S., Loidolt, P., Miltner, A., Harasek, M., [15] European Committee for Standardization CEN, Friedl, A., Production of micro- and nanoscale 2006a, Environmental management – Life cycle lignin from wheat straw using different precipi- assessment – Principles and framework (ISO tation setups. Molecules, 2018. 23(3), 633. 14040:2006), Brussels, Belgium. [6] Beisl, S., Biermair, F., Friedl, A., Mundigler, N., [16] European Committee for Standardization CEN, Miltner, A., Sequential Extrusion and organo- 2006b, Environmental management – Life cycle solv Pre-treatment for Wheat Straw Valorization. assessment – Requirements and guidelines (ISO Chemical Engineering Transactions, 2017. 61: 14044:2006), Brussels, Belgium. p. 853-858. [17] Riepl, S., Economic evaluation of innovative bio- [7] National Institute of Standards, Chemistry Web- refinery scenarios. Master thesis, 2018, TU Wien. Book, NIST Standard Reference Database Num- [18] Branker, K., Pathak, M., Pearce, J., A review of ber 69. http://webbook.nist.gov/. solar photovoltaic levelized cost of electricity. [8] Drljo, A., Simulation of biorefinery platforms for Renewable and Sustainable Energy Reviews, combined biohydrogen and biogas production 2011, 15: 4470-4482. from wheat straw, PhD Thesis, 2016, TU Wien. [19] Weinwurm, F., Fractionation and Degradation [9] Weinwurm, F., Drljo, A., Waldmüller, W., Fiala, of Wheat Straw during Liquid Hot Water and B., Niedermayer, J., Friedl, A., Lignin concen- Ethanol organosolv Treatments. PhD thesis, tration and fractionation from ethanol organosolv 2016, TU Wien. liquors by ultra- and nanofiltration, J. Cleaner [20] García, A., Alriols, M. G., Llano-Ponte, R., Production, 2016. 136: p. 62-71 Labidi, J., Energy and economic assessment of [10] Drljo, A., Liebmann, B., Wukovits, W., Friedl, soda and organosolv biorefinery processes, 20th A., Predicting minimum energy conditions for a European Symposium on Computer Aided distillation column by design of experiments and Process Engineering – ESCAPE 20, Ischia, process simulation, Chemical Engineering Trans- Naples, Italy, 06.-09.06.2010. actions, 2012. 29: p. 325-330. [21] Viell, J., Harwardt, A., Seiler, J., Marquard, W., [11] Drljo, A., Modarresi, A., Weinwurm, F., Wuko- Is biomass fractionation by organosolv-like pro- vits, W., Friedl, A., Integration of an organosolv cesses economically viable? A conceptual design process applying pinch and exery analysis, 7th study. Bioresource Technology 2013. 150: p.89-97. International Conference on Environmental En- [22] Harasek, M., Haddadi, B., Jordan, C., Friedl, A. gineering and Management – ICEEM07, Vienna, Proceedings of the 28th European Symposium on Austria, 18.-21.09.2013 in: Conference Abstracts Computer Aided Process Engineering, 2018, 1583- Book, 2013, ISBN: 978-973-621-418-9; S. 197-198. 1588. DOI: 10.1016/B978-0-444-64235-6.50276-X.

26 LENZINGER BERICHTE 94 (2018) 27 – 31

FTIR Micro-Spectroscopy for Characterisation of Radial Substituent Diffusion in Cellulose II-Beads

Danuta Mozdyniewicz1, Andreas Schwaighofer2, Karin Wieland2, Martin Häubl3, and Bernhard Lendl2

1 Kompetenzzentrum Holz GmbH , Altenberger Str.69, 4040 Linz, Austria 2 Institute of Chemical Technologies and Analytics, Vienna University of Technology, Vienna, Austria 3 Lenzing AG, Werkstraße 2, 4860 Lenzing, Austria Contact: [email protected]

Abstract

In this paper, a FTIR-based method for localization of substituents along the radius of a cellulosic bead is described. In our experiments, cellulosic beads prepared using the lyocell process are initially functionalized with citric acid (CA) and 1-octenylsuccinylanhydride (OSA), as both reagents contain IR active carboxylic groups. The analysis using FTIR micro-spectroscopy enables the determination of the analyte distribution in a simple way and accordingly reveals the diffusion ability of certain functionalizing agents. The effect of the reagent size and structure as well as reaction conditions on the penetration depth within the spheres is made visible. The maximum diffusion depth of the bulky OSA is around 400 m whereas CA is found more than twice as deep within the bead. The applied method offers the possibility of tracking the functionalization agents over the cross-section of cellulosic materials and may be applied to other forms, e. g. fibers.

Keywords: Cellulose II beads, functionalization, citric acid, 1-octenylsuccinylanhydride, FTIR Micro-Spectroscopy, radial diffusion

Introduction

The almost unlimited availability of cellulose, the ted degradation. Enhancing the surface hydrophobici- variety of processing possibilities and several advanced ty may slow down or prevent the corrosion of the applications result in a growing interest in utilization cellulosic structure. Another possibility to increase of spherical cellulosic particles in the last few years. the stability is cross-linking. The formation of intra- Cellulosic particles find application in chemistry, bio- molecular bonds decreases the swelling ability and as chemistry, chemical engineering and other related a result increases material strength [5, 6]. Epichloro- areas [1]. The three free hydroxyl groups of the glucose hydrin is the most common agent for the purpose. units making up the cellulose chain are particularly However, a substitution of the chemical is desired, accessible for functionalization, thereby enabling due to its toxic and cancerogenic properties. Many chemical modification and adaptation of the basic other suitable chemicals show similar problematic material properties. Interest in modification of cellulose properties, e. g. glutardialdehyde or glycoldiglycidyl surface has existed for a long time, and has even in- ether. The use of citric acid as an alternative cross- creased in recent years [2-4]. A main disadvantage linking agent proves ideal, due to its nontoxicity and of cellulose as a raw material is its sensitivity to pH availability [7, 8]. values deviating from neutral. The strong swelling For bead production, direct bead formation and simul- ability due to the hydrophilic character favors the dif- taneous cellulose regeneration analogous to fiber fusion into the structures and may result in accelera- spinning appears most advantageous. In this approach,

27 LENZINGER BERICHTE 94 (2018) 27 – 31

continuous production is possible and additionally the Unless otherwise specified, all the chemicals used properties may be changed at any given moment. were obtained by commercial suppliers (purity grade Therefore, procedures can be easily adapted that are at least lab quality). 1-octenylsuccinylanhydride (OSA), already used for filament formation. Particularly, the commonly used as a surface active agent, was kindly viscose and the lyocell processes seem appropriate provided by Miliken Chemicals. due to their wide spread industrial application [9, 10]. Other less common methods, e.g. dissolution in ionic Functionalized Bead Preparation liquids or NaOH/urea may be used as well, but are of For preparation of the cross-linked beads, a solution less interest due to non-industrial scale [11, 12]. containing 19.23 g citric acid (CA) and 5.62 g of Depending on the manufacturing method, different KH2PO4 was prepared in deionized H2O in a 50 ml variations of the resulting bead structure may be ob- volumetric flask. 0.98 g of the beads were suspended served. Applying the lyocell process, particles with a in 14 ml of the solution and left unstirred for 30 minutes distinctive shell-core structure are generated, imply- to swell. Subsequently, the beads were filtered off and ing the localization of the functionalization agent pri- dried at 160 °C in vacuum to start the cross-linking marily on the surface [13]. Employing functionaliza- reaction, due to ester formation of the citric acid and tion reagents with different degrees of hydrophilicity the cellulose hydroxyl groups. After 15 minutes, the leads to varying diffusion paths within the spheres. As material was washed with warm deionized water and a result, different substituent distribution within the air dried. cellulose bead may occur. Therefore, determination of For functionalization with OSA, an initial solvent the substitution pattern within the structure is essen- exchange was necessary to remove water, as a non- tial for the pinpointing of potential applications. aqueous solvent was necessary for the reaction. Fourier transform infrared (FTIR) spectroscopy is a Therefore, beads with 6.44 g dry cellulose content nondestructive, versatile, and label-free technique for were suspended in acetone that was replaced by ace- chemical analysis of a wide range of samples. It offers tone once and twice by toluene. After the solvent ex- attractive features such as fast sample preparation, change procedure the beads were stirred in 250 mL sound sensitivity, and the capacity for comprehensive toluene containing 15 mL pyridine at 60 °C for one qualitative as well as accurate quantitative analysis. hour under argon atmosphere. Subsequently, 31.74 g This spectroscopic technique has been successfully OSA were dropwise added and the suspension was applied for the characterization of cellulosic fibers further stirred at 90 °C for 18 hours under argon. After [14-17]. the reaction time, the functionalized beads were In this work, cellulosic beads manufactured by dissolv- filtered, washed with acetone and air dried [18]. ing pulp in N-methylmorpholine-N-oxide (NMMO) and regenerating the so produced dope were used as Sample Preparation for Analysis raw material for functionalization. Alcenyl succinic Air dried cellulose beads that were functionalized anhydrides are commonly used sizing agents increas- with citric acid (CA) and 1-octenylsuccinylanhydride ing the hydrophobicity of paper, whereas citric acid is (OSA) were cryo-cut at approx. -15°C to expose the a small hydrophilic molecule enabling cellulose cross- center of the bead so that the cross section could be linking due to two reactive groups. Application of investigated by FTIR microscopy (Figure 1). The ob- the two different functionalization agents allowed the tained hemispheres were glued to a microscope slide. investigation of diffusion depending on the reagent properties and the treatment conditions. The analyses of the treated beads were performed using FTIR microscopy. Materials & Methods

Materials Figure 1: Cellulose beads before (a) and after (b) the func- Cellulose II beads were obtained by Lenzing AG and tionalization reactions; (b) left: a cut bead prepared for were produced by regeneration of cellulose in water, FTIR analysis. using a device that enabled the formation of spherical particles during the regeneration step. Previously, the FTIR Microscopy pulp was dissolved in NMMO [13]. The obtained FTIR microscopic measurements were performed using beads with a diameter of ca. 2.5 mm, were washed a Bruker Hyperion 3000 IR microscope (Bruker, Ett- NMMO free and air dried. lingen, Germany) equipped with a liquid nitrogen cooled

28 LENZINGER BERICHTE 94 (2018) 27 – 31

HgCdTe (mercury cadmium telluride, MCT) detector, tion of suitable functionalization agents. On that account, connected to a Bruker Tensor 37 FTIR spectrometer. rather low mass increases were observed, suggesting Attenuated total reflection (ATR) microscopy spectra fairly low degrees of substitution (Table 1). Never- were acquired using a Ge-ATR 20x objective. For all theless, detectable bands in the FTIR spectra were measurements, spectra were recorded as the co-addi- sufficient to indicate a successful functionalization and tion of 100 scans with a spectral resolution of 4 cm-1. enabled to create diffusion profiles of the reagents. Instrument control and spectrum acquisition were performed with OPUS 7.2 software (Bruker, Ettlin- Table 1: Yields after functionalization reactions gen, Germany). In one measurement run, IR micro- Functionalization agent minitial mfinal Δ m Δ m scopic measurements of the bead cross section were [g]* [g]* [g] [%] performed beginning at the center and continuing Citric acid (CA) 0.98 1.03 0.05 5.1 outwards with a distance between measurements of 1-Octenylsuccinylanhydride (OSA) 6.44 6.48 0.04 0.6 approx. 80 µm (Figure 2). After every measurement * o.d. beads run, the Ge-ATR crystal was cleaned with water and ethanol. FTIR microscopy was performed on the bead cross For evaluation of the penetration depth of the substrates, sections revealed by cryo-cutting. ATR-IR microscopic areas of the substrate bands were normalized with an spectra were measured at a distance of approx. 80 µm, IR band attributed to the cellulose chain. The obtained resulting in 15-16 measurement points across the bead ratios were then normalized to values between 0 and radius. On one cross section, multiple measurement 1. These normalization steps were necessary to account runs were performed (Figure 2). The IR spectra of the for differences in the IR absorbance due to unequal untreated and modified cellulose beads are shown in pressure and contact between the ATR crystal and the Figure 3 and show characteristic bands of cellulosic cellulose bead surface. material [16]. Furthermore, the modified beads feature IR bands at 1600 cm-1 (OSA) and 1580 cm-1 (CA), which have been attributed to the carboxylate group of functionalized cellulose [19, 20]. The IR spectra were analysed for evaluation of the penetration depth of the substrates into the cellulosic beads. Figure 4 shows the normalized IR intensities of the IR bands attributed to OSA and CA. Absorbance values of OSA rapidly decline within the exterior 400 µm of the cellulose beads. In contrast, IR band intensities of CA drop slowly until reaching a plateau at approx. 1000 µm within the cel- Figure 2: left: Schematic of measurement procedure. ATR-IR lulose bead. The general decrease of substrate moieties microscopic measurements were performed on the bead cross towards the bead center seems reasonable, since deriva- section, starting in the center. right: Picture of the bead cross tization of the hydroxyl group within the bead interior section after performing multiple measurement runs. The is limited by diffusion, while the outmost hydroxyl indents of the ATR crystal are visible on the bead surface. groups of the beads are readily accessible. The relative density of the functional groups, however, differs be- Results & Discussion tween OSA and CA. In this context, the C8 group of OSA may limit further diffusion into the bead and this In this paper, cellulosic beads after cross-linking with substrate reacts within the outer layers of the bead. In citric acid and 1-octenylsuccinylanhydride are de- contrast, the sterically more compact CA diffuses and scribed. In both investigated cases, a change of color reacts further into the cellulose bead, even though was observed after the reaction completion, which this reagent possesses two reactive groups for cross- may indicate a surface functionalization as well as the linking. formation of cellulose degradation products and other As citric acid could be found throughout the entire co-products and their adsorption at the surface. Yellow- bead, it could be used to enhance the stability of the ing of cellulosic materials after a citric acid treatment cellulosic spheres. The cross-linking affects the cellu- has been described in the literature before [7]. The ex- lose hydroxyl groups and may prevent degradation periments were performed once and no optimization reactions to a certain extent, as the cellulose molecules of the reaction conditions was made, since the work are connected after the reaction resulting in decreased was part of a project dealing with screening and evalua- accessibility [21]. A similar effect on wood was

29 LENZINGER BERICHTE 94 (2018) 27 – 31

demonstrated by Bishof Vucusic [8]. The stability of Acknowledgements wood samples after CA treatment was enhanced due to prevented swelling, further increasing the micro-tensile Financial support was provided by the Austrian strength. These results suggest a stabilizing effect on government, the provinces of Lower Austria, Upper the treated beads, however, additional experiments Austria, and Carinthia as well as by Lenzing AG. We would be necessary to confirm the assumption. also express our gratitude to the Johannes Kepler Uni- versity, Linz, the University of Natural Resources and Life Sciences (BOKU), Vienna, and Lenzing AG for their in-kind contributions. A.S., K.W. and B.L. gratefully acknowledge financial support by the Austrian research funding association (FFG) under the scope of the COMET programme within the research project “Industrial Methods for Process Analytical Chemistry – From Measurement Technologies to Information Systems (imPACts)” (contract #843546).

References

[1] Gericke, M., J. Trygg, and P. Fadim, Functional Figure 3: IR spectra of cellulose beads (black) without func- Cellulose Beads: Praparation, characterization, tionalization, and modification with (blue) OSA and (red) CA. and applications. Chem Rev, 2013. 113: p. 4812- 4836. [2] Rogowin, S.A., Neue Erfolge auf dem Gebiet der chemischen Modifizierung von Zellulose. Len- zinger Ber, 1970. 30: p. 16-25. [3] Dimow, K., et al., Iionenaustauschende Fasern auf Cellulosebasis. Lenzinger Ber, 1978. 45: p. 60-65. [4] Mais, U., et al., Functionalization of cellulose. Lenzinger Ber, 2000. 79: p. 71-76. [5] Morales, A., et al., Adsorption and releasing properties of bead cellulose. Chinese J Polym Sci, 2004. 22(5): p. 417-423. Figure 4: Normalized IR intensities of characteristic bands [6] Ishimura, D., Y. Morimoto, and S. Hidenao, In- of (blue) OSA and (red) CA plotted versus the penetration fluences of chemical modifications on theme- depth within the cellulsose bead. chanical strength of cellulose beads. Cellulose, 1998. 5(2): p. 131-151. Conclusions [7] Xun, H. and X. Zhou, Mechanism of citric acid crosslinking in cotton fabrics. J China Textile In this report we demonstrate for the first time that University, 1999. 16(2): p. 16-20. FTIR microscopy is an excellent method for determi- [8] Bischof Vuksic, S., et al., Polycarboxylic acids nation of functionalization patterns within cellulose as non-formaldehyde anti-swelling agents for beads of sufficient size, providing the presence of IR- wood. Holzforschung, 2006. 60: p. 439-444. active groups in the reagent. In our case, the influence [8] Weber, H., Overview on latest developments in of the reagent structure and size on the penetration viscose industry. Lenzinger Ber, 2005. 84: p. 8-16. depth was shown. Citric acid was detected almost [10] Fink, H.P., et al., Structure formation of regene- throughout the entire bead, whereas the highest con- rated cellulose materials from NMMO-solutions. centrations of 1-octenylsuccinylanhydride were found Progress in polymer science, 2001. 26(9): p. closer to the surface. 1473-1524.

30 LENZINGER BERICHTE 94 (2018) 27 – 31

[11] Trygg, J., et al., Physicochemical design of the [17] Fan, M., D. Dai, and B. Huang, Fourier Trans- morphology and ultrastructure of cellulose be- form Infrared Spectroscopy for Natural Fibres, ads. Carbohyd Polym, 2013. 93: p. 291-299. in Fourier Transform – Materials Analysis, S.M. [12] De Oliveira, W. and W.G. Glasser, Hydrogels Salih, Editor. 2012, InTech: China. from polysaccharides. II. Beads with cellulose [18] Belhalfaoui, B., et al., Succinate-bonded cellulo- derivatives. Journal of Applied Polymer Science, se: A regenerable and powerful sorbent for cad- 1996. 61(1): p. 81-86. mium-removal from spiked high-hardness ground- [13] Häubl, M., J. Innerlohinger, and C. Schirk, water. J Hazard Mater, 2009. 169: p. 831-837. Three-dimenstional cellulose moulded body, [19] McSweeny, J.D., R.M. Rowell, and S.-H. Min, mothod for the production and use of the same. Effect of Citric Acid Modification of Aspen 2015. WO2015054711 (A2). Wood on Sorption of Copper Ion. J Nat Fibers, [14] Garside, P. and P. Wyeth, Identification of cellu- 2006. 3(1): p. 43-58. losic fibres by FTIR spectroscopy. Stud Conserv, [20] Quellmalz, A. and A. Mihranyan, Citric Acid 2003. 48: p. 269-275. Cross-Linked Nanocellulose-Based Paper for [15] Carillo, A., et al., Structural FTIR analysis and Size-Exclusion Nanofiltration. ACS Biomater thermal characterisation of lyocell and viscose Sci Eng, 2015. 1(4): p. 271-276. type fibres. Eur Polym J, 2004. 40: p. 2229-2234. [21] Schramm, C. and B. Rinderer, Application of Po- [16] Comnea-Stancu, I.R., et al., On the identification lycarboxylic Acids to Cotton Fabrics under mild of Rayon/Viscose as a Major Fraction of Micro- cruing Conditions. Institute of Textile Chemistry plastics in the Marine Environment: Discrimina- and Textile Physics, 2006. 40(9-10): p. 799- 804. tion between Natural and Man-made Cellulosic Fibers by Fourier Transform Infrared Spectro- scopy. Appl Spectrosc, 2017. 71(5): p. 939-950.

31 LENZINGER BERICHTE 94 (2018) 33 – 42

Key Aspects of the Bleaching of Industrially Found Cellulosic Chromophores

Nele Sophie Zwirchmayr1, Ute Henniges1,4, Thomas Röder2, Antje Potthast1, and Thomas Rosenau1,3*

1 Division of Chemistry of Renewable Resources, Department of Chemistry, University of Natural Resources and Life Sciences, Muthgasse 18, A-1190 Vienna, Austria 2 Lenzing AG, Werkstraße 2, 4860 Lenzing, Austria 3 Johan Gadolin Process Chemistry Centre, Åbo Akademi University, Porthansgatan 3, Åbo/Turku FI-20500, Finland 4 Conservation of Works of Art on Paper, Archival and Library Materials, State Academy of Art and Design Stuttgart, Höhenstraße 16, 70736 Fellbach, Germany * Contact: [email protected]

Abstract

The cellulosic chromophores 2,5-dihydroxy-[1,4]-benzoquinone, 5,8-hydroxy-[1,4]-naphthoquinone, 2,5-dihydroxyacetophenone and 2,6-dihydroxyacetophenone, main contributors to cellulose discoloration, were bleached with H2O2, O3 and ClO2 at conditions resembling the corresponding sequence in industrial pulp bleaching (P, Z, or D-stage). Measurement of reaction kinetics allowed conclusions on activation energy, reaction mechanism, activation parameters and the presence of colored intermediates. Final degradation products were determined; small carboxylic acids form the major part of them.

Keywords: cellulosic chromophores, P-stage, Z-stage, D-stage, 2,5-dihydroxy-[1,4]-benzoquinone, 5,8-hydroxy-[1,4]-naphthoquinone, hydroxyacetophenones

Introduction

As a million tons per year industry, the pulp and paper and further oxidation leads to additional chromophoric industry is a key player in European economies and compounds. Recently, the link between the chromo- holds the potential for the development of environ- phore precursor hexeneuronic acid (4-deoxy--L- mentally benign and sustainable consumer products. threo-hex-4-enopyranosiduronic acid, HexA) found For the industry highest product brightness is the goal, in glucuronoxylan-containing cellulosic pulps (hard- ensured by extensive bleaching. Cellulose-based pro- wood) and cellulosic key chromophores was estab- ducts are also supposed to stay white over time and lished [8,9,10]. show as little ageing (yellowing) as possible. However, To reveal the nature of chromophores commonly found maximum brightness comes with a downside: oxidative, in pulp and paper, the chromophore release and identi- hydrolytic and thermal stress inflicted on the poly- fication method (CRI) was developed. CRI is therefore saccharide raw material, caused by the extensive use of an analytical method specifically targeting cellulosic bleaching chemicals. The connection between damages chromophores. The CRI method is based on applica- in the polysaccharide and the formation of colored tion of a Lewis acid and a reductant (BF3*2 HOAc/ molecules that reduce brightness was established in Na2SO3) by which chromophores can be released from the seventies by Theander and his co-workers. The pulp and separated by common analytical techniques. group described how chemical stress can cause ali- Typically, these chromophores are present in remarkably phatic carbohydrates to form aromatic structures even low amounts which range from ppm to ppb. Although without the presence of any lignin [1,2,3,4,5,6,7]. present in such small amounts, their high extinction Some of these aromatics have a yellow tint themselves, coefficients in the VIS region is the reason they give a

33 LENZINGER BERICHTE 94 (2018) 33 – 42

yellow tint to cellulose products that is immediately quinones, and hydroxy-acetophenones. Within these recognizable by the human eye. The CRI method has three classes, three structures were isolated that been applied to a manifold of cellulose products, among together make up about 80% of the total chromophore them lignin free cellulosic materials, such as aged cotton mass (Scheme 1): 2,5-dihydroxy-[1,4]-benzoquinone or aged bacterial cellulose. CRI revealed that cellu- (1), 5,8-hydroxy-[1,4]-naphthoquinone (2), and 2,5-di- losic chromophores belong to three compound classes: hydroxyacetophenone (3). These three compounds are hydroxy-[1,4]-benzoquinones, hydroxy-[1,4]-naphtho- thus regarded as key chromophores [11,12,15].

2 H2O -2 H2O

Scheme 1: The key chromophores 2,5-dihydroxy-[1,4]-benzoquinone (1), 5,8-hydroxy-[1,4]-naphthoquinone (2), and 2,5-dihydroxyacetophenone (3). Hydrogen bonds present in solid state and acidic media are formed between carbonyl oxygens and the H-atom in OH moieties are marked by dashed lines. In alkaline media, deprotonation results in the anions 1a from 1, 2a from 2 and 3a-3e from 3.

Chromophores found in pulp and paper can be sub- side is the large amount of sample needed and its in- divided into primary chromophores and secondary ability to be automated. Improvements in the analysis chromophores. A primary chromophore is formed from of cellulosic chromophores have been made by imple- the polysaccharide by oxidative and hydrolytical stress menting the ambient ionization techniques DESI (de- during bleaching or aging (see Scheme 1); in secondary sorption electrospray ionization), PS (paper spray), and chromophores, however, part of the solvent or process derivatization methods simpler than the CRI method chemicals are found incorporated into the chromophore. that allow detection by GC-ECD (electron capture de- From that we can conclude that secondary chromophores tection) and EI-GC-MS (electron ionization) with limits are dependent on the underlying process; for example, of detection as low as 60 ng chromophore/g cellulose in chromophores found in the spinning bath of viscose [16,17]. solutions, sulfur atoms originating from CS2 can be Besides their strong color, further important charac- found, as well as motifs from the solvent NMMO, or teristics of cellulosic chromophores are their high chlorine atoms after a D-stage; see Scheme 2 for details stabilization that roots in H-boding in acid and solid [13,14,15,53]. state and resonance stabilization in alkaline media Although the CRI method allowed to identify the (Scheme 1) [18,19,20]. The latter impedes an attack at most common cellulosic key chromophore, its down- a double bond, the mode of action of many oxidizers.

34 LENZINGER BERICHTE 94 (2018) 33 – 42

Scheme 2: Various types of secondary chromophores where functional groups of solvent molecules can be found incorporated into the chromophores (CS2 or NMMO). The same is possible for the bleaching chemical, as in the presented case of ClO2 bleaching.

Bleaching studies using the neat chromophores in set- ness. Bleaching studies were performed with the ups mimicking industrial bleaching are a powerful key chromophores DHBQ, DHNQ, 2,5-DHAP, and tool to optimize bleaching sequences and the use of 2,6-DHAP (4) and the most used bleaching agents bleaching chemicals, and to achieve the economically H2O2, O3 and ClO2. An overview of the bleaching and environmentally most sustainable product bright- scheme is given in Figure 1.

Figure 1: Overview of the bleaching experiments performed with the chromophores 1-4 and the bleaching agents H2O2, O3, and ClO2.

Results and discussion H2O2 bleaching with large excesses of peroxide (see Table 1 for details) are presented together with Arrhe- Peroxide bleaching nius plots in Figure 2 [23]. Arrhenius plots are con- Hydrogen peroxide (H2O2) is applied in P-stages in structed from temperature variation while keeping the industrial bleaching. Among its advantages are its easy peroxide content constant and are needed to calculate storability and its harmless decomposition products, activation energy Ea, activation parameters and the water and oxygen. When H2O2 bleaching, correspond- reaction constant k. Ea and k are crucial when inves- ing to an industrial P-stage was applied to the chromo- tigating the bleaching behavior of chromophores. phores 1-4, the observed kinetics were of pseudo first Ea represents how “hindered” or “approachable” a order throughout (linear correlation of ln[c] vs. time) chemical reaction is, as it is the energy barrier that has [23,21,22]. Typical degradation curves obtained from to be overcome in order for the reaction to take place.

35 LENZINGER BERICHTE 94 (2018) 33 – 42

As a rule of thumb, chemical reactions with activation industrial bleaching we should keep in mind that the energies >25 kcal/mol are not likely to take place at higher the temperature and the higher k, the faster the room temperature and would need catalysis or heating reaction; a fact that we also observed in our H2O2 of the reaction mixture. In our case, the observed acti- bleaching experiments. vation energies were below this threshold. The reac- Reaction conditions, activation parameters, activation tion rate constant k, on the other hand, is a parameter energy Ea(exp) as calculated from Arrhenius plots of that represents the speed at which a chemical reac- the peroxide bleaching of DHBQ, DHNQ, 2,5-DHAP, tion takes place. It is temperature dependent, thus for and 2,6-DHAP are presented in Table 1.

Figure 2: H2O2 degradation of DHNQ. Left: ln(A) vs. time with various excesses of peroxide. The dashed line represent the linear fit. Right: Arrhenius plot constructed from the data obtained by temperature variation from 30 °C to 70 °C while keeping the excess peroxide constant at 100 mol H2O2 / 1 mol DHNQ. k-values correspond to the temperatures 30-70 °C. Graphs adapted from reference [23].

Table 1: Parameters of bleaching experiments performed with the chromophores DHBQ, DHNQ, 2,5-DHAP, and 2,6-DHAP and H2O2 [23,21].

Chromophore Molar excess of peroxide Activation energy k [s-1] and relative to chromophore [kcal / mol] temperature range [°C] DHBQ 257 16.1 0.00064-0.0261 5-50 DHNQ 5-667 11.7 0.00396-0.0351 30-70 2,5-DHAP 100 -1100 8.79 0.0011-0.0060 36-80 2,6-DHAP 100 -1100 10.60 0.0034-0.0245 36-80

Ozone bleaching viscosity. This process is dependent on the pH applied Ozone is used increasingly in the pulp and paper in- and other factors, such as temperature and metal ions dustry. Although ozone is harmful for most organisms, present. These chemical reactions can be used to one’s the environmental concerns over its use are still lower advantage, as they allow to tailor the viscosity of a than in the case of chlorine-based bleaching agents. cellulose product according to purpose [24,41,42,43, That is mainly due to its harmless degradation product 44,45,46,47]. The reactivity of ozone towards poly- oxygen. Ozone’s reaction with double bonds has been saccharide-derived key chromophores 1-4 was in- the subject of many studies. First suggestions on the vestigated by dissolving them at pH 2 with the aid of mechanism by Criegee were followed by extensive co-solvent (MeOH), sampling before and during the computational and analytical studies on the complexes bleaching and subsequent VIS measurements. Degra- and ozonides formed during ozonolysis [25,26,27,28, dation products were analyzed by GC-MS and NMR 29,30,31,32,33,34,35,36,37,38,39,40]. In pulp bleach- measurements. The ozone load was 0.021 g O3/l and ing, the formation of different radical species and the gas flow of 50 l/h, bleaching was performed at other oxidative species, for example •OH and H2O2, room temperature. An overview of the bleaching con- have to be taken into account. Radicals like •OH ditions applied is given in Table 2. can damage the polysaccharide and result in lower

36 LENZINGER BERICHTE 94 (2018) 33 – 42

Table 2: Chromophore amounts, solvent composition and volumes of chromophore solutions used in ozone bleaching experiments, as well as total bleaching time until decoloration.

Chromo- Amount Concentration H2O (pH 2, MeOH Total phore [mg] [mmol/l] H2SO4) [ml] [ml] bleaching time [s] 1 184 4.378 290 10 480 2 52 0.912 200 100 900 3 187 4.083 200 100 360 4 185 4.026 300 50 480

Ozone bleaching results were promising: Although in there is no defined reaction order immediately deriv- some cases the formation of a colored intermediate able from the graphs. In this case, the activation energy was observed (DHNQ, 2,6-DHAP), all chromophores Ea cannot be calculated from VIS measurements. Al- could be fully degraded – see Figure 3 for degradation though the activation energies of O3 could not be cal- curves obtained from VIS measurements. As visible culated, it was possible to compare the chromophores from the curves of the O3 degradation of chromophores, among each other, as presented in Table 3.

Figure 3: Kinetic analyses from ozone bleaching experiments of the cellulosic chromophores DHBQ, DHNQ, 2,5-DHAP, and 2,6-DHAP. The absorption at 400 nm is displayed and in case of DHNQ additionally also at 517 nm where the formation of a colored intermediate was observed in the full spectra from 200 to 700 nm.

Table 3: Comparison of the time until full decoloration of chromophore solutions in ozone bleaching.

Chromophore t [s] Chromophore t [s] 2-OH-NQ 203 DHBQ 386 2,5-HAP 293 DHNQ 1843 2,6-HAP 367

37 LENZINGER BERICHTE 94 (2018) 33 – 42

DHNQ and DHBQ showed to be the most resistant tion reaction. Colored intermediates are found in most to ozone bleaching, an observation that matches their bleaching reactions, although sometimes we can only chemical behavior: in H2O2, these two chromophores speculated about their nature. had the highest activation energies (16.1 and 11.7 kcal / mol, respectively) [21,23]; their high thermodynamic Colored intermediates stability is discussed in references 18 and 19. In ozone bleaching, VIS measurements revealed the formation of colored intermediates for some chromo- Chlorine dioxide phores (DHNQ and 2,6-DHAP); similar observation Although chlorine dioxide is a very strong oxidant were made in H2O2 and ClO2 bleaching. These inter- and known for its high reactivity [48,49,50,51,52] and mediates can have VIS absorptions comparable to or selectivity – it leaves the polysaccharide largely higher than the parent chromophore. Their structure is unharmed but targets lignin – in pulp bleaching its be- not always known, and usually, colored intermediates havior towards chromophores showed to be less are degradable by the bleaching agent that initially “straightforward” than for example it is the case with caused their formation; however, the fact that they H2O2. When the reactivity of ClO2 with the key chro- exist also emphasizes the importance of at least 200 mophore DHNQ was assessed, VIS measurements molar excesses in peroxide bleaching and > 5 molar revealed that DHNQ is re-formed in the case of the excessed for ClO2 (in case of O3 no such description reaction of ClO2 molar equivalents (referring to moles can be made) and also points out the importance of DHNQ) are less than 5. A peculiar behavior is also adequate mixing during the bleaching process. Ex- reported from ClO2 and DHBQ, where several colored amples for such colored intermediates are presented intermediates are formed in the course of the oxida- in Scheme 3.

Scheme 3: Quinones 5, 6 and 7 from 2,6-DHAP and 2,5-DHAP from peroxide bleaching, respectively; tetrahydroxy benzoquinone (THBQ) and rhodizonic acid (RA) from the ClO2 bleaching of DHBQ, 1,4,5,8-naphthalenetetrone (8) from P-stage bleaching of DHNQ as examples for chromophore intermediates formed during pulp bleaching [53].

Final degradation products dation products were colorless. The substances were found in chromophore bleaching in good agreement with literature reports on similar When bleaching chromophore solutions, the application experiments. Additionally, it should be pointed out of H2O2, O3, or ClO2 results in large similar degradation that the small amounts of carboxylic acids generated products, which are the gaseous CO2 and carboxylic from chromophore degradation should not cause any acids. Table 4 gives and overview of the found degra- fluctuation of pH, scaling or sedimentation. If that dation products and the bleaching methods applied. is the case during industrial processes, other causes We were pleased to observe that, besides the forma- than the chromophore degradation should be con- tion of colored intermediates, all found final degra- sidered.

38 LENZINGER BERICHTE 94 (2018) 33 – 42

Table 4: Overview of carboxylic acids found in chromophore bleaching as determined by GC-MS and NMR measurements of the reaction mixtures.

39 LENZINGER BERICHTE 94 (2018) 33 – 42

Conclusions in slightly acidic, aqueous solution. Acta Chem. Scand., Ser. B 1978, B32 (4), 249-56. Experimental investigations of chromophore bleach- [5] Theander, O.; Westerlund, E., Formation of aro- ing were performed under conditions similar to indus- matic compounds from carbohydrates. VIII. Re- trial bleaching using H2O2, O3, and ClO2 and the action of D-erythrose in slightly acidic, aqueous cellulosic chromophores DHBQ, DHNQ, 2,5-DHAP, solution. Acta Chem. Scand., Ser. B 1980, B 34 and 2,6-DHAP. In all cases it was possible to fully (10), 701-5. degrade the chromophores, but colored intermediates [6] Theander, O.; Nelson, D. A.; Hallen, R. T., Forma- were observed in most bleaching reactions. To avoid tion of aromatic compounds from carbohydrates. chromophore (re)formation, the molecular excess of the X. Reaction of xylose, glucose, and glucuronic bleaching agent should be around 200 molar equivalents acid in acidic solution at 300°C. Prepr. Pap. – referring to chromophore for H2O2 bleaching and at Am. Chem. Soc., Div. Fuel Chem. 1987, 32 (2), least 5 molar equivalents in the case of ClO2 bleaching 143-7. (for O3 no values can be given) and the importance of [7] Nelson, D. A.; Hallen, R. T.; Theander, O., Forma- homogeneity and adequate mixing is pointed out. We tion of aromatic compounds from carbohydrates. observed that for H2O2 temperatures above room tem- Reaction of xylose, glucose, and glucuronic acid perature – we investigated up to 80 °C – are beneficial in acidic solution at 300°. ACS Symp. Ser. 1988, for the bleaching process. O3 bleaching was perfor- 376 (Pyrolysis Oils Biomass), 113-18. med in solution at room temperature. Literature states [8] Rosenau, T.; Potthast, A.; Zwirchmayr, N. S.; than O3 bleaching benefits from lower temperatures Hettegger, H.; Plasser, F.; Hosoya, T.; Bacher, due to the reduced formation of competing radical M.; Krainz, K.; Dietz, T., Chromophores from that damage the polysaccharide, but the temperature hexeneuronic acids: identification of HexA-derived applied industrially certainly has to be a compromise chromophores. Cellulose (Dordrecht, Neth.) 2017, of desired molecular weight of the final product and 24 (9), 3671-3687. energy efficiency, as cooling and heating are energy [9] Rosenau, T.; Potthast, A.; Zwirchmayr, N. S.; demanding processes. Finally, all observed degradati- Hosoya, T.; Hettegger, H.; Bacher, M.; Krainz, on products were colorless carboxylic acids or CO2. K.; Yoneda, Y.; Dietz, T., Chromophores from hexeneuronic acids (HexA): synthesis of model Acknowledgements compounds and primary degradation intermediates. Cellulose (Dordrecht, Neth.) 2017, 24 (9), The authors would like to thank the Austrian Research 3703-3723. promotion Society (FFG, project 829443) for finan- [10] Zwirchmayr, N. S.; Hosoya, T.; Hettegger, H.; cial support. Bacher, M.; Krainz, K.; Dietz, T.; Henniges, U.; Potthast, A.; Rosenau, T., Chromophores from References hexeneuronic acids: chemical behavior under peroxide bleaching conditions. Cellulose (Dordrecht, [1] Popoff, T.; Theander, O., Formation of aromatic Neth.) 2017, 24 (9), 3689-3702. compounds from carbohydrates. I. Reaction of [11] Rosenau, T.; Potthast, A.; Krainz, K.; Hettegger, D-glucuronic acid, D-galacturonic acid, D-xylose, H.; Henniges, U.; Yoneda, Y.; Rohrer, C.; French, and L-arabinosein slightly acidic, aqueous solu- A. D., Chromophores in cellulosics, XI: isolation tion. Carbohyd. Res. 1972, 22 (1), 135-49. and identification of residual chromophores from [2] Popoff, T.; Theander, O., Formation of aromatic bacterial cellulose. Cellulose (Dordrecht, Neth.) compounds from carbohydrates. Part III. Reac- 2014, 21 (4), 2271-2283. tion of D-glucose and D-fructose in slightly acidic, [12] Rosenau, T.; Potthast, A.; Milacher, W.; Hofinger, aqueous solution. Acta Chem. Scand., Ser. B A.; Kosma, P., Isolation and identification of 1976, B30 (5), 397-402. residual chromophores in cellulosic materials. [3] Olsson, K.; Pernemalm, P. A.; Popoff, T.; Thean- Polymer 2004, 45 (19), 6437-6443. der, O., Formation of aromatic compounds from [13] Adorjan, I.; Potthast, A.; Rosenau, T.; Sixta, H.; carbohydrates. V. Reaction of D-glucose and Kosma, P., Discoloration of cellulose solutions methylamine in slightly acidic, aqueous solution. in N-methylmorpholine-N-oxide (Lyocell). Part Acta Chem. Scand., Ser. B 1977, B31 (6), 469-74. 1: Studies on model compounds and pulps. Cel- [4] Olsson, K.; Pernemalm, P. A.; Theander, O., For- lulose (Dordrecht, Neth.) 2005, 12 (1), 51-57. mation of aromatic compounds from carbo- [14] Rosenau, T.; Potthast, A.; Milacher, W.; Adorjan, hydrates. VII. Reaction of D-glucose and glycine I.; Hofinger, A.; Kosma, P., Discoloration of cel-

40 LENZINGER BERICHTE 94 (2018) 33 – 42

lulose solutions in N-methylmorpholine-N-oxide [25] Gillies, J. Z.; Gillies, C. W.; Lovas, F. J.; Matsu- (Lyocell). Part 2: Isolation and identification of mura, K.; Suenram, R. D.; Kraka, E.; Cremer, D., chromophores. Cellulose (Dordrecht, Neth.) 2005, Van der Waals complexes of chemically reactive 12 (2), 197-208. gases: ozone-acetylene. Journal of the American [15] Korntner, P.; Hosoya, T.; Dietz, T.; Eibinger, K.; Chemical Society 1991, 113 (17), 6408-6415. Reiter, H.; Spitzbart, M.; Roeder, T.; Borgards, [26] Gillies, C. W.; Gillies, J. Z.; Suenram, R. D.; A.; Kreiner, W.; Mahler, A. K.; Winter, H.; Lovas, F. J.; Kraka, E.; Cremer, D., Van der Groiss, Y.; French, A. D.; Henniges, U.; Potthast, Waals complexes in 1,3-dipolar cycloaddition A.; Rosenau, T., Chromophores in lignin-free reactions: ozone-ethylene. Journal of the Ameri- cellulosic materials belong to three compound can Chemical Society 1991, 113 (7), 2412-2421. classes. Chromophores in cellulosics, XII. Cel- [27] Epstein, S. A.; Donahue, N. M., Ozonolysis of Cyc- lulose (Dordrecht, Neth.) 2015, 22 (2), 1053- lic Alkenes as Surrogates for Biogenic Terpenes: 1062. Primary Ozonide Formation and Decomposition. [16] Schedl, A.; Korntner, P.; Zweckmair, T.; Henni- J. Phys. Chem. A 2010, 114 (28), 7509-7515. ges, U.; Rosenau, T.; Potthast, A., Detection of [28] Chuong, B.; Zhang, J.; Donahue, N. M., Cyclo- Cellulose-Derived Chromophores by Ambient alkene Ozonolysis: Collisionally Mediated Ionization-MS. Anal. Chem. (Washington, DC, Mechanistic Branching. Journal of the American U. S.) 2016, 88 (2), 1253-1258 Chemical Society 2004, 126 (39), 12363-12373. [17] Schedl, A.; Zweckmair, T.; Kikul, F.; Bacher, M.; [29] Vayner, G.; Addepalli, S. V.; Song, K.; Hase, W. L., Rosenau, T.; Potthast, A., Pushing the limits: Post-transition state dynamics for propene Quantification of chromophores in real-world ozonolysis: Intramolecular and unimolecular paper samples by GC-ECD and EI-GC-MS. dynamics of molozonide. J. Chem. Phys. 2006, Talanta 2018, 179, 693-699. 125 (1), 014317/1-014317/16. [18] Hosoya, T.; French, A. D.; Rosenau, T., Chemistry [30] Wheeler, S. E.; Ess, D. H.; Houk, K. N., Think- of 5,8-dihydroxy-[1,4]-naphthoquinone, a key ing Out of the Black Box: Accurate Barrier chromophore in aged cellulosics. Mini-Rev. Org. Heights of 1,3-Dipolar Cycloadditions of Ozone Chem. 2013, 10 (3), 302-308. with Acetylene and Ethylene. J. Phys. Chem. A [19] Hosoya, T.; French, A. D.; Rosenau, T., Chemistry 2008, 112 (8), 1798-1807. of 2,5-dihydroxy-[1,4]-benzoquinone, a key [31] Kuczkowski, R. L., The structure and mecha- chromophore in aged cellulosics. Mini-Rev. Org. nism of formation of ozonides. Chemical Society Chem. 2013, 10 (3), 309-315. Reviews 1992, 21 (1), 79-83. [20] Hosoya, T.; French, A. D.; Rosenau, T., Chemistry [32] Bernatek, E.; Moeskeland, J.; Valen, K., Ozonoly- of 2,5-dihydroxyacetophenone, a Key Chromo- sis of phenols. II. Pyrocatechol, resorcinol, and phore in Cellulosic Substrates. Mini-Rev. Org. quinol. Acta Chem. Scand. 1961, 15, 1454-60. Chem. 2015, 12 (1), 88-95. [33] Bernatek, E., Ozonolyses in the Naphthoquinone [21] Hosoya, T.; Rosenau, T., Degradation of 2,5- and Benzofuran Series. Oslo Univ. Press: 1960; Dihydroxy-1,4-benzoquinone by Hydrogen p 144 pp. Peroxide under Moderately Alkaline Conditions [34] Bernatek, E., Ozonolysis of naphthoquinones. I. Resembling Pulp Bleaching: A Combined Kinetic Ozonolysis of 1,4-naphthoquinone in chloroform. and Computational Study. J. Org. Chem. 2013, Tetrahedron 1958, 4, 213-22. 78 (22), 11194-11203 [35] Bernatek, E., The ozonization of p-benzoquinone. [22] Hosoya, T.; Rosenau, T., Degradation of 2,5- Acta Chem. Scand. 1954, 8, 1943. Dihydroxy-1,4-benzoquinone by Hydrogen [36] Bernatek, E.; Straumsgard, K. A., Ozonolysis of Peroxide: A Combined Kinetic and Theoretical p-benzoquinone. II. Acta Chem. Scand. 1959, Study. J. Org. Chem. 2013, 78 (7), 3176-3182. 13, 178-85. [23] Zwirchmayr, N. S.; Hosoya, T.; Henniges, U.; [37] Pillar, E. A.; Guzman, M. I., Oxidation of Subs- Gille, L.; Bacher, M.; Furtmueller, P.; Rosenau, tituted Catechols at the Air-Water Interface: Pro- T., Degradation of the Cellulosic Key Chromo- duction of Carboxylic Acids, Quinones, and Poly- phore 5,8-Dihydroxy-[1,4]-naphthoquinone by phenols. Environmental Science & Technology Hydrogen Peroxide under Alkaline Conditions. 2017, 51 (9), 4951-4959. J. Org. Chem. 2017, 82 (21), 11558-11565. [38] Geletneky, C.; Berger, S., The Mechanism of [24] Ragnar, M. On the Importance of Radical For- Ozonolysis Revisited by 17O-NMR Spectroscopy. mation in Ozone Bleaching, Kungliga Tekniska European Journal of Organic Chemistry 1998, Högskolan, 2000. 1998 (8), 1625-1627.

41 LENZINGER BERICHTE 94 (2018) 33 – 42

[39] Criegee, R., Mechanism of Ozonolysis. Ange- [47] Chirat, C.; Lachenal, D., Other ways to use ozone wandte Chemie International Edition in English in a bleaching sequence. Tappi J. 1997, 80 (9), 1975, 14 (11), 745-752. 209-214. [40] Ragnar, M.; Eriksson, T.; Reitberger, T.; Brandt, [48] Leigh, J. K.; Rajput, J.; Richardson, D. E. Kine- P., A new mechanism in the ozone reaction with tics and Mechanism of Styrene Epoxidation by lignin-like structures. Holzforschung 1999, 53 Chlorite: Role of Chlorine Dioxide. Inorg. Chem. (4), 423-428. 2014, 53, 6715-6727. [41] Pi, Y.; Zhang, L.; Wang, J., The formation and [49] Aguilar, C. A. H.; Narayanan, J.; Singh, N.; influence of hydrogen peroxide during ozonation Thangarasu, P. Kinetics and mechanism for the of para-chlorophenol. J. Hazard. Mater. 2007, oxidation of anilines by ClO2: a combined expe- 141 (3), 707-712. rimental and computational study.J. Phys. Org. [42] Buehler, R. E.; Staehelin, J.; Hoigne, J., Ozone Chem. 2014, 27, 440-449. decomposition in water studied by pulse radioly- [50] Lehtimaa, T.; Kuitunen, S.; Tarvo, V.; Vuorinen, sis. 1. Perhydroxyl (HO2)/hyperoxide (O2-) and T., Reactions of aldehydes with chlorous acid HO3/O3- as intermediates. The Journal of Physi- and chlorite in chlorine dioxide bleaching. Holz- cal Chemistry 1984, 88 (12), 2560-2564. forschung 2010, 64, 555-561. [43] Gierer, J., Formation and involvement of super- [51] Napolitano, M. J.; Green, B. J.; Nicoson, J. S.; oxide (O2·-/HO2.) and hydroxyl (OH·) radicals in Margerum, D. W., Chlorine Dioxide Oxidations TCF bleaching processes. A review. Holzfor- of Tyrosine, N-Acetyltyrosine, and Dopa.Chem schung 1997, 51 (1), 34-46. Res Toxicol 2005, 18, 501-508. [44] Chirat, C.; Lachenal, D., Effect of Hydroxyl Radi- [52] Hull, L. A.; Davis, G. T.; Rosenblatt, D. H.; Wil- cals in Ozone Bleaching. Holzforschung 1997, liams, H. K. R.; Weglein, R. C., Oxidations of 51 (2), 147-154. amines. III. Duality of mechanism in the reac- [45] Pouyet, F.; Lachenal, D.; Das, S.; Chirat, C., tion of amines with chlorine dioxide. J. Am. Minimizing viscosity loss during totally chlorine- Chem. Soc. 1967, 89, 1163-70. free bleaching of hardwood kraft pulp. Bio [53] Hettegger, H.; Hosoya, T.; Rosenau, T., Che- Resources 2013, 8 (1), 238-249, 12 pp. mistry of the Redox Series from Hexahydroxy- [46] Chirat, C.; Lachenal, D., Effect of ozone on pulp benzene to Cyclohexanehexanone. Curr. Org. components. Application to bleaching of kraft Synth. 2016, 13 (1), 86-100. pulps. Holzforschung 1994, 48 (Suppl.), 133-9.

42 LENZINGER BERICHTE 94 (2018) 43 – 55

Cellulose Carboxylate/Tosylate Mixed Esters: Dependence of their Physicochemical Properties on the Degree of Carboxylate Substitution

Daniela C. Ferreira,1,2 Gustavo S. Bastos,1 Annett Pfeiffer,3 Denise F. S. Petri,2 Thomas Heinze,3* and Omar A. El Seoud2*

1 Institute of Chemistry, the University of São Paulo, Prof. Lineu Prestes Av., 748, 05508-000, São Paulo, SP; * Contact: [email protected] 2 Institute for Technological Research of the State of São Paulo S.A., Laboratory of Pulp and Paper, Prof. Almeida Prado Av., 532, 05508-901, São Paulo, SP, Brazil 3 Centre of Excellence for Polysaccharide Research, Institute of Organic and Macromolecular Chemistry, Friedrich Schiller University of Jena, Humboldtstraße 10, D-07743, Jena, Germany; * Contact: [email protected]

Abstract

The dependence of the physicochemical properties of cellulose carboxylate/tosylate mixed esters on their molecular structures is reported. The esters employed have a constant degree of substitution (DS) of the tosyl group, DSTosyl = 0.98 and variable DS of the ester group, DSAcyl (ester = acetate, butanoate, hexanoate). These mixed esters do not exhibit glass transition temperatures and decompose before melting. Using perichromic dyes and contact angle measurements of their films, the dependence of the following properties on DSAcyl were investigated: empirical polarity, Lewis acidity and basicity, and hydrophobicity. Most of these properties display a simple dependence on DSAcyl, which reflects the conversion of the hydroxyl- into ester group. These mixed esters were regenerated as uniform microspheres from their solutions in acetone; their adsorption of methylene blue and Sudan IV dyes was studied, and the dye adsorption data correlate with the Lewis basicity and the dispersive component of surface energy.

Keywords: cellulose mixed ester; cellulose carboxylate/tosylate, thermal analysis, perichromic indicator, contact angle; microspheres, dye absorption.

Introduction

The impetus for studying cellulose derivatives with Most research was carried out on mixed esters of C2-C4 mixed functionalities is that their properties are dif- carboxylic acids. Cellulose carboxylate/tosylate mixed ferent, sometimes better compared to the corresponding esters (Cell-carboxylate/tosylate) carry strongly dipo- derivatives with one functionality. For example, mixed lar groups (C=O and S=O) with different basicities carboxylic esters, e.g., cellulose acetate/propionate, and a polarizable aromatic ring. Additionally, intro- acetate/butanoate, and acetate/phthalate have proper- duction of sulphur as a substituent into cellulose de- ties (luster, hardness) better than simple esters, e.g., creases the decomposition temperature of the derivative, cellulose acetate, propionate and phthalate, probably and suppresses its flaming combustion by reducing because of the juxtaposition of both acyl moieties in the amount of flammable volatile products produced the same anhydro-glucose unit (AGU). Cellulose mixed on thermal degradation [4]. esters are employed in sheeting, molding plastics, film It is expected that the presence of different functionali- products, lacquer and melt dip coatings, and controlled ties and the control of the degree of substitution (DS) drug release [1-3]. allows the modulation of physicochemical properties

43 LENZINGER BERICHTE 94 (2018) 43 – 55

of Cell-carboxylate/tosylate. This increases their dye, and regenerated them form acetone solutions as potential use in applications where acid-base and hydro- microspheres with narrow size distribution [7]. phobic interactions are important, e.g., immobilization In the present work, we examined in details some of biomacromolecules by adsorption. The fact that the physicochemical properties of these esters that are tosylate is a good leaving group allows the preparation relevant to applications. In particular, we wanted to of novel deoxy-cellulose derivatives containing halo- correlate the above-mentioned properties with the gen atoms [5] or amino groups [2]. Moreover, some of structural variable across the series, i.e., DSAcyl, and these mixed esters can be processed by extrusion from between the three esters series (acetates, butanoates, and the melt; form films by casting from their solutions in hexanoantes) at comparable DSAcyl. We determined organic solvents; and microspheres by precipitation in Lewis acidity and basicity of their films by using peri- water [6,7]. chromic dyes, vide infra, and extracted information Recently, we synthesized a series of Cell-carboxylate/ about film hydrophobicity and surface energy from tosylate mixed esters by acylation of cellulose tosylate contact angle measurements. All examined properties with carboxylic acid anhydrides (acetic-, butanoic- and showed simple dependence (linear or not) on DSAcyl. hexanoic anhydride) as shown in Figure 1. The starting The properties of the films are affected by the acid- cellulose tosylate had a (fixed) DSTosyl = 0.98 and the base affinity of carbonyl (RCO-), modulated by steric final mixed esters hadAc DS = 0.37-1.50 (acetates), effects of the alkyl group (RCO-). We found that dye DSBu = 0.39-0.92 (butanoates), and DSHx = 0.26-1.05 adsorption efficiency by microspheres of these mixed (hexanoate). We casted these mixed esters as films; esters is related to their Lewis basicity and the dis- determined their empirical polarity by a perichromic persive component of surface energy.

Figure 1: Scheme of cellulose acetate/tosylate synthesis: (i) microcrystalline cellulose (MCC) dissolution in N,N-dimethylacetamide (DMAc)/LiCl; (ii) tosylation of cellulose, and (iii) acetylation of cellulose tosylate, where R = H in unsubstituted hydroxyl or CH3CO in cellulose acetate/tosylate. We depict: 1 = cellulose tosylate; 2 = cellulose deoxy-chloro (side product DSCl = 0.03 to 0.07), and 3 = cellulose acetate/tosylate [7].

Experimental

Chemicals The chemicals were purchased from Sigma-Aldrich or (DTBSB), 1-methyl-5-nitroindoline (MNI) and 2-(N,N- Merck and were purified as recommended elsewhere dimethylamino)-7-nitro-9-H-fluorene (DMANF) were [8]. The compounds 5-nitroindoline (NI); -carotene, available from previous studies [9,10]. The molecular methylene blue (MB), and Sudan IV (SU-IV) are com- structure of the dyes employed in the present work is mercial. The perichromic probes 2,6-dichloro-4-(2,4,6- shown in Figure 2. triphenylpyridium-1yl) (WB), o-tert-butylstilbazoilum betaine (TBSB), o,o´-di-tert-butylstilbazolium betaine

44 LENZINGER BERICHTE 94 (2018) 43 – 55

Figure 2: Molecular structures of perichromic probes and dyes employed in the present work. (A) 2,6-dichloro-4-(2,4,6-triphenyl- pyridinium-1-yl)phenolate (WB); (B) o-tert-butylstilbazoilum betaine (TBSB); (C) o,o´-di-tert-butylstilbazolium betaine (DTBSB); (D) 5-nitroindoline (NI); (E) 1-methyl-5-nitroindoline (MNI); (F) 2-(N,N-dimethylamine)-7-nitro-9H-fluorene (DMANF); (G) -carotene; (H) methylene blue and (MB); (I) Sudan IV (SU-IV).

Cellulose Carboxylate/ ped with model IST-204A reflectance accessory; TGA Tosylate Mixed Esters curves, TA instruments Q50 thermo-balance; DSC The cellulose carboxylate/tosylate mixed esters that we curves, TA Instruments Q20; TG-MS, Netzsch STA employed were those available from our previous study, 490 PC Luxx thermo-analyzer coupled to Aëlos QMS see Figure 1 [7]. 403C mass-spectrometer; contact angle, Surface Electro Optics, model Phoenix-i. Equipments The following equipment were employed: UV-VIS reflectance spectra of cellulose mixed ester films, Shimadzu UV-2550, UV-VIS spectrophotometer equip-

45 LENZINGER BERICHTE 94 (2018) 43 – 55

Measurements Contact Angle and Surface Energy of Cell-Carbo- xylate/Tosylate Films Thermal Analysis We prepared Cell-carboxylate/tosylate films by casting To obtain the DSC curves of cellulose mixed esters, we from their solutions (0.1 g/mL) in DMSO on clean weighted ca. 3 mg of each sample in a sealable alumi- silicon dishes, and evaporated the solvent under reduced num crucible; the latter was sealed and heated at 10 K/ pressure at 40 °C to allow the formation of smooth min, under a dynamic nitrogen atmosphere (100 mL/ films [12]. min). First, we annealed each sample by heating from In order to determine the surface energy of the poly- 25 °C to 220 °C; cooled it to 25 °C, then heated it again. mers (γs), contact angle (θ) measurements were per- The DSC cell was calibrated by using Indium (m.p. formed with diiodomethane (θDIM) and distilled water 156.6 oC; Hfusion = 28.42 J/g) and Zinc (m.p. 419.6 (θW) at (24 ± 1) °C by the sessile drop method [13], oC; Hfusion = 112.0 J/g). the droplet volume used for the measurements was 10 For recording the thermogravimetric curves we heated µL. At least three samples of the same material were ca. 5 mg sample contained in a platinum crucible from analyzed. 25 °C to 400 °C, at heating rate of 10 K/min, under a The surface energy (γs) of the cellulose mixed ester p d dynamic nitrogen atmosphere, 50 mL/min. We calcu- films and its polar (γs ) and dispersive (γs ) compo- lated the decomposition temperature (TDecomp) from nents were calculated using the Owens-Wendt's equa- the first derivative of thermogravimetric curve, and tion [14], also known as the geometric mean equation. p analyzed the gases liberated during sample decompo- Values of θW and θDIM, were measured and the γs d sition by mass spectrometry. and γs components of the surface tension of diiodo- -3 2 p -3 2 methane (γL = 50.8 x 10 J/m ; γL = 50.8 x 10 J/m ) -3 2 p -3 2 Determination of Surface Properties of Cellulose and water (γL = 72.8 x 10 J/m ; γL = 51.0 x 10 J/m d -3 2 Mixed Esters Films by Perichromic Probes and γ L = 21.8 x 10 J/m ) were taken from the litera- Films of cellulose tosylate and Cell-carboxylate/tosylate ture [15]. containing the perichromic dyes shown in Figure 2 The advancing (θa) and receding (θr) contact angles were prepared as describe previously [7]. Briefly, cel- were measured at (24 ± 1) °C for water droplets. θa lulose mixed ester solutions (0.1 g/mL) were prepared and θr were determined employing droplet volume of in acetone containing the appropriate perichromic 10 L and 5 L, respectively. The contact angle hys- probe (concentration = 10-3 mol/L for probes B, C teresis (Δθ) was calculated as the difference between -4 and F, 10 mol/L for probes D, E and G, depicted in θa and θr values [13]. At least three samples of the Figure 2). We prepared films of cellulose tosylate and same material were analyzed. Cell-carboxylate/tosylate manually by drop casting We employed one-way analysis of variance (ANOVA) on clean glass plates, followed by solvent evaporation with post hoc (p) test to evaluate the differences of under reduced pressure at room temperature. data among groups, and considered p < 0.05 as signifi- The reflectance spectra of cellulose tosylate and Cell- cant difference. carboxylate/tosylate films were recorded at room temperature, under the following conditions, against Dye Adsorption Test by Cellulose Mixed Ester BaSO4 as white reference: scanning rate = 140 nm/ Microspheres min; slit width = 1.0 nm; spectral resolution = 0.5 nm. Cell-carboxylate/tosylate samples (each with DSAcyl ≈ 0.4 The reflectance spectra were converted into the cor- and 1.0) were shaped into microspheres by drop responding absorption curves by the Kubelka-Munk technique as described elsewhere [7]. The influence function [11] as implemented in version 2.21 of of DSAcyl on dye adsorption capacity of Cell-carbo- Shimadzu UV-probe program. The values of max xylate/tosylate microspheres was tested using MB and were calculated from the first derivative of the absorp- SU-IV. tion spectra; the uncertainty in max was ± 1 nm. Cell-carboxylate/tosylate (5 mg) were shaped into The perichromic properties, empirical polarity spheres resulting in 2.5 mL of acetone/water suspen- (ET(WB)); Lewis acidity (SA), Lewis basicity (SB), sion (20% of acetone in water). To this suspension of dipolarity (SD) and polarizability (SP) were calcula- spheres, we added 0.5 mL of MB solution (430 mg/L); ted from the max of the intramolecular charge transi- stirred the suspension for 24 h (25 ± 1 °C); removed tion band of appropriated perichromic probe. The cal- microspheres by centrifugation (10 000 g), and deter- culations are shown in Supporting Information, SI. mined the UV-VIS absorbance of the supernatant. Be- cause of insolubility of SU-IV in water, we employed a protocol similar to that employed for MB, except for the solvent employed for microsphere formation and

46 LENZINGER BERICHTE 94 (2018) 43 – 55

dye adsorption (pure 2-propanol) and dye concentration 221 ± 4 °C (TDecomp2) in agreement with literature (206 mg/L). The dye removal efficiency by the micro- [17,18]. spheres was calculated as recommended elsewhere, see We gain information about the thermal decomposition the Supporting Information [16]. of these mixed esters from analysis of sample weight loss, and m/z of the gases evolved during their heat- Results and Discussion ing; these results are depicted in Figure SI 2 and Table 1. The relevant information is that during TDecomp1 we Note: The calculations of all cellulose mixed ester pa- observed peaks at m/z =18 (water) and 44 (CO2), cha- rameters are given in Supporting Information. racteristic of cellulose decomposition, and the peak at m/z = 60 due to McLafferty rearrangement of the acyl Dependence of the Thermal Properties moiety. During TDecomp2 we observed peaks at m/z = + of Cell-Carboxylate/Tosylate on DSACYL 64 (SO2) and m/z = 91 (tropylium cation, C7H7 ), attributed to the degradation of the sulfonate moiety. Thermagravimetric Analysis For all Cell-carboxylate/tosylate mixed esters, how- Typical TGA curves of cellulose tosylate and Cell- ever, the characteristic peak of 4-toluenesulfonic acid carboxylate/tosylate with DSAcyl ≈ 1.0 and their para- at m/z = 173 was not observed. Our mass spectroscopy meters are shown in Figure SI 1A and Table SI 1. All data indicate, therefore, that the first reactions that TGA curves showed two-stage weight loss starting occurs during thermal decomposition of the mixed for the mixed esters at 185 ± 3 °C (TDecomp1) and at esters is their deacylation.

Table 1: Values of m/z and the corresponding fragments produced by thermal decomposition of cellulose tosylate and Cell-carboxylate/tosylate

The influence of DSAcyl on decomposition temperatures Differential Scanning Calorimetry (DSC) of cellulose mixed esters is shown in Figure 3. It was Typical DSC curves of cellulose tosylate and Cell- found that the two initial decomposition temperatures carboxylate/tosylate are shown in Figure SI 1B, one (IDT) of Cell-carboxylate/tosylate show a simple de- exothermic and one endothermic peaks (Table SI 2) pendence on DSAcyl; IDT increase with the increasing are presented in curves of all samples. The maximum DSAcyl. The results of the corresponding regression temperature of exothermic peak (Texo) of all samples analysis are listed in Table 2. The dependence of IDT- shows a linear correlation with TDecomp1 (Figure 3C, 1 on DSAcyl is an ascending exponential as observed entry 8-10 in Table 2), indicating that the exothermic for pure cellulose acetate [9], which is in agreement peak correspond to the biopolymer decomposition, in with the TG-MS data that indicates the scission of agreement with the second heating DSC curve, where acyl units during the first decomposition stage. no events are observed Figure SI 1C.

47 LENZINGER BERICHTE 94 (2018) 43 – 55

Figure 3: Dependence of the initial decomposition temperatures (IDT) on DSAcyl of Cell-carboxylate/tosylate. (A) First decomposition stage and (B) second decomposition stage. (C) Correlation between DSC and TGA data.

In summary, we conclude the following about the ing correlation coefficients. A version of Equation (1), thermal properties of Cell-carboxylate/tosylate: (i) applied for a perichromic empirical polarity probe, substitution of the hydroxyl groups of AGU by carbo- e.g., WB (ET(WB)) is: xylic ester group produces increase in the corresponding decomposition temperatures; these are correlated ET(WB) = a SA + b SB + d SD + p SP with DSAcyl by simple equations; (ii) scission of the (2) acyl group occurs first; (iii) the glass transition- and where ET(WB) is the surface empirical polarity. melting temperatures of these mixed esters are higher than their decomposition temperatures, as is observed We are interested in understanding and quantifying for cellulose proper. solute-cellulose derivative interactions. Previously we studied cellulose carboxylic esters [9] and cellulose Influence of DSACYL on Surface Proper- ethers [21]. To our knowledge, this is the first study on ties of Cell-Carboyxylate/Tosylate Mixed mixed esters. Esters: Use of Perichromic Probes and The effects of the structure of the mixed esters on the Contact Angle Measurement properties of the corresponding films (Lewis acidity, Lewis basicity, etc.) are due to the presence of both A successful approach to describe the effect of medi- tosyl and acyl moieties. We attribute the response of um on chemical phenomena is to use a linear combi- the perichromic dyes to the latter moiety because nation of medium descriptors, e.g., as shown in Equa- DSTosyl is constant through the series. Previous data on tion (1) [19, 20]. the same mixed esters showed that the effects of the tosyl and carboxyl moieties are independent. For example, Effects of the medium = a SA + b SB + d SD + p SP ῦ(C=O) and ῦ(asSO2) are linear functions in DSAcyl (1) [7]. We assume that the same independent contribu- where, SA, SB, SD, and SP refers to Lewis acidity, tions to the perichromic parameters are operative for Lewis basicity, dipolarity, and polarizability of the carboxylate/tosylate esters. In other words, we expected medium, respectively; a, b, d, and p are the correspond- a correlation, linear or not, between DSAcyl and

48 LENZINGER BERICHTE 94 (2018) 43 – 55

Table 2: Dependence of the thermal properties of Cell-carboxylate/tosylate on the degree of substitution of carboxylate moieties

a) IDT = initial temperature of decomposition. 1 and 2 correspond, respectively, to the first and second decomposition stage observed on TGA curve of Cell-carboxylate/tosylate. b) TDecomp1 = temperature of the first stage of decomposition, determined from DTG curves. c) Texo = temperature of exothermic peak present in DSC curves.

ET(WB). We also expected that application of the by increased hydrophobic interactions and steric hin- multi-parameter Equation (2) indicates which film drance because the preferential conformation of property controls its polarity; a useful information for RCO-, e.g., perpendicular or tilted toward the film in- applications of these mixed properties, e.g., in enzyme terface [24]. A similar line of reasoning applies to ana- immobilization. lysis of the left-hand terms of Equation (2). The effects of RCO- on the perichromic parameters are due to dipolar and acid/base interactions of RCO-, ET(WB) as well as hydrophobic and steric interactions of Previously, was showed that ET(WB) of Cell-carbo- RCO-. As a model for the former interactions we use xylate/tosylate films decrease linearly as a function of ET(WB) of ethyl acetate, ethyl butanoate, and ethyl increasing DSAcyl; the slopes of the corresponding hexanoate: 46.5; 45.3; 44.6 kcal/mol, respectively straight lines increase on going from acetate to hexa- [22]. The molar volumes (cm3/mol) of the parent car- noate (Entries 1 to 3 in Table 3) [7]. Consequently, for boxylic acids: 56.1 ± 3.0; 89.1 ± 3.0; 122.2 ± 3.0 [23] the same mixed ester series, this decrease is a conse- for acetic, butyric and hexanoic acid, respectively, can be quence of replacement of the strongly dipolar OH used as model for evaluation of the steric interactions. group by the weaker dipolar RCO-. Applying the results of these model compounds to the present cellulose mixed esters, we expect a decrease in ET(WB) in going from the cellulose acetate/tosylate to cellulose hexanoate/tosylate, probably modulated

49 LENZINGER BERICHTE 94 (2018) 43 – 55

Table 3: Dependence of the perichromic properties of Cell-carboxylate/tosylate on the degree of ester substitution

Lewis Acidity and Lewis Basicity from the substitution of the hydroxyl groups, hydrogen The values calculated of SA and SB are listed in Table bond donor, by acyl moiety, hydrogen bond acceptor. SI-3 and are plotted in Figure 4 and 5. The depen- The Lewis acidity dependence on RCO- shows the dence of SA on DSAcyl is linear (Entries 4 to 6 in following order: acetyl > hexanoyl > butanoyl reflec- Table 3). For the same series of mixed esters, the de- ting the above-mentioned acid-base character of crease of SA as a function of increasing DSAcyl results RCO-, probably coupled with steric effects of RCO-.

Figure 4: Dependence of Lewis acidity of cellulose carboxylate/tosylate films on the degree of substitution of carboxylate unit. Parts A, B and C refer to acetate/tosylate, butanoate/tosylate and hexanoate/tosylate, respectively. For comparison with the acetate/tosylate series, we calculated the dependence of SA on DSAcyl > unity for the butanoate/tosylate and hexanoate/tosylate- shown in part D- from the regression equations of Table 3.

50 LENZINGER BERICHTE 94 (2018) 43 – 55

The interplay between acid-base character of RCO- length of RCO- on SB is evident only after the mini- and steric effects of RCO- is also evident in the corre- mum is reached, with the following order: cellulose lation of SB with DSAcyl (Figure 5), being a second butanoate/tosylate > cellulose hexanoate/tosylate > order polynomial with minimum at DSAcyl ≈ 0.8 for all cellulose acetate/tosylate; again reflecting the inter- mixed ester series, (Entries 7 to 9 in Table 3). As can play between acid/base character and steric interac- be seen from part D of Figure 5, the effect of the tions.

Figure 5: Dependence of Lewis basicity of cellulose carboxylate tosylate films on degree of substitution of carboxylate unit. Parts A, B and C refer to cellulose acetate/tosylate, cellulose butanoate/tosylate and cellulose hexanoate/tosylate, respectively.

For comparison with the acetate/tosylate series, we calculated the dependence of SB on DSAcyl > unity for the butanoate/tosylate and hexanoate/tosylate – shown in part D – from the regression equations of Table 3.

Dipolarity and Polarizability will be similarly affected because we use max of The remaining medium descriptors of Equation (2) -carotene for calculation of SD, as given elsewhere are SP and SD. It was not feasible to calculate these [26]. Consequently, we thought information regarding descriptors from the spectra of the perichromic probes these two descriptors from contact angle measure- -carotene and 2-(N,N-dimethylamine)-7-nitrofluorene ments as indicated below. (Figure 2), as indicated elsewhere [25, 26]. The reason is the small value SP (= SP cellulose acetate/tosylate – Contact Angle and Surface Energy SP cellulose hexanoate/tosylate), ca. 0.03 at DSAcyl ≈ The contact angle (θ) of a liquid droplet on a surface 1.0). This small variation agrees with the SP = 0.04, results from the equilibrium of three forces, namely, calculated for model compounds ethyl acetate and the liquid–air (γL), liquid–solid (γSL), and solid–air ethyl hexanoate [26]. Consequently, the variation of (γS) interfacial tensions, as proposed by Young [27]: SP as a function of DSAcyl, or the length of RCO- is within experimental uncertainty. The correlation of γS= γSL+ γL cosθ (3) film dipolarity (SD) with the same structural variables

51 LENZINGER BERICHTE 94 (2018) 43 – 55

The work of adhesion (WSL) between the liquid and acyl groups to surface properties of Cell-carboxylate/ the surface can be determined by the Young-Dupré tosylate films. equation [27-29]: The θW values determined for water are a direct measurement of films hydrophobicity; the larger the θW WSL= γL (1+cosθ) (4) value, the more hydrophobic is the surface. The values are summarized in Table 4. Our data show that the The Owens-Wendt geometric mean equation allows hydrophobicity of Cell-carboxylate/tosylate decreased calculation of WSL from the geometric mean of polar in the following order: cellulose hexanoate/tosylate > (γp) and dispersive (γd) components of the surface and cellulose butanoate/tosylate > cellulose acetate/tosylate. liquid: This result agrees with the expected effect of increasing the length of (hydrophobic) RCO-. With few ex- ceptions, an increase of DSAcyl for the same mixed ester series results in an increase θW , due to the sub- d p Generally two liquids with known γL and γL compo- stitution of OH by RCO. d p nents are used for the contact angle measurements and The γs and γs components of the surface energy (γS) d p (γs ) and (γs ) determination. Those parameters indicate of Cell-carboxylate/tosylate films (Table 4) were cal- the contribution of each type of interaction (dipolar- culated from the values of θW and θDIM values using d p or van der Waals interactions) between solid and liquid Equation (5). Although the values of γS, γs and γs for p to reduce the interfacial tension. High values of γs cellulose acetate/tosylate films are higher than the other indicate that the surface is enriched by hydrophilic Cell-carboxylate/tosylate esters the difference is, how- d d groups, whereas the γs component includes forces ever, small. The contribution of γs to γS was the between permanent dipoles, permanent dipole and a largest regardless of the length of RCO. A similar re- corresponding induced dipole, and induced dipoles. sult was observed for cellulose acetate and cellulose Therefore, we employed the surface energy method as mixed carboxylic esters (cellulose acetate/propionate an alternative to the perichromic approach to evaluate and acetate/butanoate) [30]. the dipolar- and van der Waals contributions of the

Table 4: Values of θ calculated for water (W) and diiodomethane (DIM) droplets on cellulose ester films and the corresponding p d polar (γs ) and dispersive (γs ) components of the surface energy (γS)

a) The symbol (*) indicates that the data are statistically different (p < 0.05), in the corresponding data set.

52 LENZINGER BERICHTE 94 (2018) 43 – 55

p The dependence of γs on RCO- follows the order ace- dyes from solution was calculated from the UV-VIS tyl > butanoyl > hexanoyl, as expected for the contri- absorbance spectra (Table 5). Within the same mixed bution of alkyl chain of acyl groups and is opposite to ester series, the removal efficiency (R.E.) of MB in- d hydrophobicity tendency. The γs values presented no creased with increasing DSAcyl, indicating that dye-ester statistically significant variation on RCO- or DSAcyl, hydrophobic interactions are also operative. Values of in agreement with perichromic results for SP and SD. R.E. are not significantly different within the three Therefore, both techniques (contact angle and peri- series, for the same DSAcyl, in agreement to the surface chromism) are not sensitive enough to evaluate the energy data. dependence of the dispersive force on DSAcyl. In case of SU-IV, the values of R.E. are very close for We note that the homogeneity of Cell-carboxylate/ the three Cell-carboxylate/tosylate and practically in- d tosylate films was adequate for the surface energy de- dependent of DSAcyl, which is in agreement with γs termination, because the mean contact angle hystere- and SP/SD values. It is also possible that dye self-aggre- sis (∆θ) values for most samples was (16 ± 4)°, as gation, [31] contributed to this “leveling” of adsorp- shown in Table SI 4, indicating that the films presented tion within the esters, because dye-dye interactions smooth homogeneous surfaces. efficiently competes with dye-microsphere interactions. However, dye self-aggregation is suppressed in Dependence of Dye Adsorption on DSAcyl aliphatic alcohols, [32] hence stronger interaction of of Cell-Carboxylate/Tosylate dye-cellulose derivative is expected. Therefore, the As given by Equation (1), the effect of medium (cellu- low R.E.(SU-IV) values probably reflects the weak lose derivative) on a phenomenon of interest (adsorp- dispersive forces as indicated by perichromic and tion) depends on adsorbate-surface specific and non- contact angle techniques. specific interactions. As an example, we tested the ad- In summary, if MB is taken as a model for charged sorption of two distinct dyes, water-soluble, cationic adsorbates, then it is it enough to use members of one MB and oil-soluble, neutral SU-IV by these mixed carboxylate/tosylate series with different DSAcyl to in- esters. We used cellulose acetate/tosylate, cellulose duce detectable differences in solute adsorption. We butanoate/tosylate and cellulose hexanoate/tosylate intend to test the generality of the previous statement with DSAcyl ≈ 0.4 and 1.0, shaped into microspheres by carrying out additional adsorption experiments with with average diameter of (200 ± 35) nm [7]. adsorbates whose chemical structures are varied in a The efficiency of Cell-carboxylate/tosylate micro- systematic way. spheres (1 wt % of mixed ester in water) to remove the

Table 5: Dye removal efficiency of Cell-carboxylate/tosylate microspheres

53 LENZINGER BERICHTE 94 (2018) 43 – 55

Conclusions References

A series of cellulose carboxylate/tosylate mixed esters [1] K. J. Edgar, C. M. Buchanan, J. S. Debenham, with a constant DSTosyl and variable DSAcyl, with in- P. A. Rundquist, B. D. Seiler, M. C. Shelton, creasing hydrophobicity of the carboxylate moiety D. Tindall, Prog. Polym. Sci., 2001, 26, 1605. (acetate, butanoate, hexanoate) were characterized re- [2] D. Klemm, B. Heublein, H. P. Fink, A. Bohn, garding thermal behavior and surface properties. The Angew. Chem. Int. Ed., 2005, 44, 3358. cellulose carboxylate/tosylate showed two tempera- [3] K. J. Edgar, Cellulose, 2007, 14, 49-64. tures of decomposition where scission of the acyl [4] R. K. Jain, K. Lal, H. L. Bhatnagar, J. Appl. moiety occurs before that of the tosyl counterpart. Polym. Sci., 1987, 33, 247. These mixed esters did not exhibit glass transition- or [5] G. Siegmond, D. Klemm, Polym. News, 2002, melting temperatures. Consequently, they can only be 27, 84. regenerated from solution, i.e., not by extrusion from [6] S. Hornig, T. Heinze, Biomacromolecules, 2008, the melt. Surface properties, including empirical polar- 9, 1487. ity, Lewis acidity and Lewis basicity were calculated [7] D. C. Ferreira, G. S. Bastos, A. Pfeifer, T. Heinze, from the UV-Vis spectra of perichromic probes. Their O. A. El Seoud, Carbohydr. Polym., 2016, 152, 79. values are modulated by dipolar and steric effects of [8] W. L. Armarego, C. L. Chai, Purification of the RCO- group, and showed simple dependences on Laboratory Chemicals, 6th ed. Elsevier Ltd., DSAcyl. The small effects of DSAcyl on the spectra of New York, 2009. the appropriate perichromic dyes precluded calculation [9] L. C. Fidale, C. Ißbrücker, P. L. Silva, C. M. of SD and SP of the films. We sought this information Lucheti, T. Heinze, O. A. El Seoud, Cellulose, from contact-angle measurements; this technique also 2010, 17, 937. indicated the small magnitude of these two descriptors. [10] O. A. El Seoud, P. A. R. Pires, C. Loffredo, Finally, we used (water-soluble) MB and (oil-soluble) M. Imran, P. D. Pulcini, M. F. Corrêa, R. Mustafa, SU-IV to probe adsorption by cellulose carboxylate/ J. Braz. Chem. Soc., 2013, 24, 1079. tosylate microspheres. Adsorption of (ionic) MB in- [11] M. L. Myrick, M. N. Simcock, M. Baranowski, creases as a function of increasing DSAcyl, whereas H. Brooke, S. L. Morgan, J. N. McCutcheon, this structural variable has practically no effect on ad- Appl. Spectrosc. Rev., 2011, 46, 140. sorption of (nonionic) SU-IV. Adsorption efficiency [12] D. F. S. Petri, G. Wenz, P. Schunk, T. Schimmel, of both dye agrees with the results of perichromism Langmuir, 1999, 15, 4520. and contact angle measurements. [13] A. W. Adamson, Physical chemistry of surfaces, 5th ed., John Wiley & Sons Ltd., Chichester, 1990. Acknowledgements [14] D. K.Owens, R. C. Wendt, J. Appl. Polym. Sci., 1969, 13, 1741. D. C. Ferreira thanks FIPT for PDCE fellowship for [15] B. Bouali, F Ganachaud, J.-P. Chapel, C. Pichot, her stay at Friedrich Schiller University of Jena. O. A. P. Lanteri, J. Colloid Interface Sci., 1998, 208, 81. El Seoud and Denise F. S. Petri thank CNPq for re- [16] H. Deng, J. Lu, G. Li, G. Zhang, X. Wang, Chem. search productivity fellowships Grants 307022/2014-5 Eng. J., 2011, 172, 326. and 305178/2013-0, respectively; and FAPESP for [17] T. Heinze, K. Rahn, M. Jaspers, H. Berghmans, financial support (2014/22136-4). T. Heinze gratefully J. Appl. Polym. Sci., 1996, 60, 1891. acknowledged financial support from German Ministry [18] T. Heinze, K. Rahn, M. Jaspers, H. Berghmans, of Food and Agriculture based on an enactment of the Macromol. Chem. Phys., 1996, 197, 4207. German Bundestag. The authors also thanks to L. S. [19] C. Reichardt, T. Welton, Solvents and Solvent Galvão from IPT for the TGA and DSC measurements, Effects in Organic Chemistry, 4th ed., Wiley-VCH, to Prof. V. R. L. Constantino and Dr. R. A. A. de Couto New York, 2011. from IQ-USP for the TGA-MS measurements. [20] L. C. Fidale, T. Heinze, O. A. El Seoud, Carbo- hyd. Polym., 2013, 93, 129. [21] L. C. Fidale, P. M. Lima, L. M. A. Hortêncio, P. A. R. Pires, T. Heinze, O. A. El Seoud, Cellulose, 2012, 19, 151. [22] R. Casarano, L. C. Fidale, C. M. Lucheti, T. Heinze, O. A. El Seoud, Carbohyd. Polym., 2011, 83, 1285.

54 LENZINGER BERICHTE 94 (2018) 43 – 55

[23] ACD/Labs. Calculated using Advanced Chem- [28] M. K. Chaudhury, Mater. Sci. Eng., 1996, R16, 97. istry Development (ACD/Labs) Software V11.02 [29] P. M. Kosaka, J. Amim Jr, R. S. N. Saito, D. F. S. (© 1994-2017 ACD/Labs). Petri, in Model Cellulosic Surfaces – ACS Series [24] M. Wühn, J. Weckesser, C. Wöll, Langmuir, (Ed: M. Roman), American Chemical Society, 2001, 17, 7605. Washington 2009, 223. [25] C. Loffredo, P. A. R. Pires, M. Imran, O. A. El [30] P. M. Kosaka, Y. Kawano, D. F. S. Petri, J. Colloid Seoud, Dyes Pigment., 2013, 96, 16. Interface Sci., 2007, 316, 671. [26] P. A. R. Pires, M. Imran, C. Loffredo, P. M. [31] K. Hamada, H. Kubota, A. Ichimura, T. Lijima, Donate, D. Previdi, O. A. El Seoud, J. Phys. Org. S. Amiya, Ber. Bunsenges. Physik. Chem., 1985, Chem., 2013, 26, 280. 89, 859. [27] J. K. Berg, C. M. Weber, H. Riegler, Phys. Rev. [32] K. Rohatgi, G. Singhal, J. Phys. Chem., 1966, Lett., 2010, 105, 076103. 70, 1695.

55 LENZINGER BERICHTE 94 (2018) 57 – 66

Kinetics of Cellulose Acylation with Carboxylic Anhydrides and N-acylimidazoles in Ionic Liquid/Molecular Solvent Mixtures: Relevance to the Synthesis of Mixed Cellulose Esters

Thais A. Bioni,1 Naved I. Malek,2 and Omar A. El Seoud1 * 1 Institute of Chemistry, the University of São Paulo, Prof. Lineu Prestes Av., 748, 05508-000, São Paulo, Brazil. *Contact: [email protected] 2 Applied Chemistry Department, S.V. National Institute of Technology, Surat, Gujarat, India.

Abstract

Using conductivity measurements we studied the kinetics of acylation of microcrystalline cellulose (MCC) under homogeneous solution conditions in mixtures of the ionic liquid 1-allyl-3-methylimidazolium chloride (AlMeImCl) and the molecular solvents (MSs) N,N-dimethylacetamide (DMAc) and dimethyl sulfoxide (DMSO). The acylating agents were acetic- and propionic anhydride (Ac2O, Pr2O, respectively), N-acetyl-, N-benzoyl- and N-propionyl- imidazole (AcIm, BzIm, and PrIm, respectively). Third-order rate constants (k3) were calculated by dividing the observed rate constants by dividing by the concentrations of the acylating agent and AlMeImCl. We calculated the values of k3 for the reaction of MCC with each acylating agent, and for the reaction of the biopolymer with their mixtures (Ac2O+Pr2O or AcIm + BzIm). These calculations showed that acylation by mixtures of reagents are essentially independent. The dependence of k3 on the structure of the acylating agent and the nature of the MS were explained by considering the activation parameters. We synthesized the mixed esters under the conditions of the kinetic experiments and used 1H NMR spectroscopy to calculate the partial degrees of substitution

(DS) in each case, e.g., DSAcetate and DSPropionate. The ratio between the individual k3, e.g., k3,Ac2O/k3,Pr2O and that of the corresponding partial degree of substitution, e.g., DSAcetate/DSPropionate were found to be similar. That is, kinetic data can be successfully employed to predict the product partial degrees of substitution of the mixed esters, if steric effects are not dominant, as in case of the formation of cellulose acetate/benzoate.

Keywords: kinetics, cellulose mixed esters, ionic liquid, molecular solvents

Introduction

Ionic liquids (ILs) dissolve cellulose; the resulting so- nitrile, N,N-dimethylacetamide, DMAc, and dimethyl lutions are used for the regeneration and shaping of sulfoxide, DMSO [1-3]. The use of MSs as diluents the biopolymer and its derivatization under homo- increases heat-transfer, and results in a massive de- geneous reaction conditions. Depending on the con- crease in biopolymer solution viscosity and the energy centration and characteristics of the dissolved cellulose, of viscus flow. According to the Stokes-Einstein diffu- the molecular structure of the IL, and the temperature sion equation, lower viscosity leads to higher diffusion the solutions obtained are anisotropic and very viscous; rates of the species present in solution, which corres- leading to gel formation at still higher biopolymer ponds to an increase in mass transfer, hence increase concentration. These problems can be attenuated/elimi- in reaction rate. This expectation was demonstrated nated by use of molecular solvents (MSs) as diluents experimentally, e.g., for Diels-Alder reactions in pure for the IL, especially dipolar aprotic ones, e.g., aceto- ILs, [4,5] and their mixtures with MSs [6].

57 LENZINGER BERICHTE 94 (2018) 57 – 66

The concentration of the MS in the binary solvent mix- aliphatic acids (acetic and propionic) and aliphatic ture affects important experimental variables, includ- plus aromatic acid (acetic, benzoic, propionic,). Mixed ing the concentration of dissolved cellulose and the esters with different partial degree of substitution physico-chemical properties of the resulting biopoly- (DS) of each acyl moiety, in particular cellulose ace- mer solution. For example, it was shown than the wt% tate/propionate and acetate/butanoate have favorable of dissolved cellulose in IL-DMSO and IL-DMAc properties not exhibited by the esters of a single acyl shows maxima as a function of the MS mole fraction group, probably because of the juxtaposition of both [7]. The resulting bioploymer solutions were studied acyl moieties in the same anhydroglucose unit, AGU. by solvatochromic probes. The calculated solvato- These mixed esters are employed commercially on a chromic parameters were found to depend on the large scale, e.g., in coating industry, molding plastics, nature of the IL and MS, and solution composition. For film products, lacquer- and melt dip coatings [16,17]. example, solutions of microcrystalline cellulose (MCC) Industrially, mixed esters are prepared under hetero- in IL-MS are more polar and more basic in IL-DMSO geneous conditions [18, 19]; few studies reported their than in IL-sulfolane [8]. Favorable solvent properties, synthesis under homogeneous conditions [1,20,21]. e.g., higher polarity and basicity favor cellulose disso- To our knowledge, there are no kinetic data on cellu- lution because of the disruption of the intermolecular lose acylation to form mixed esters under homogene- hydrogen bonding between biopolymer chains. The ous reaction conditions. same properties favor reactions where there is an in- In the present study, we used conductivity measure- crease in polar character between the reactants and ments to investigate the simultaneous acylation of dis- transition state, e.g., acyl transfers [9]. solved MCC by two acylating agents in mixtures of the A rational use of binary solvent mixtures, however, is IL 1-allyl-3-methylimidazolium chloride (AlMeImCl) not straightforward. The reason is that the dependence with DMAc or DMSO. The acylating agents were of the properties of the binary mixture on its compo- acetic- and propionic anhydride (Ac2O and Pr2O res- sition is not ideal, i.e., is not linear. This non-ideality pectively); N-acetyl-, N-benzoyl- and N-propionyli- applies to macroscopic properties, e.g., density and midazole (AcIm, BzIm, and PrIm, respectively), see viscosity, [11, 12] and microscopic ones, e.g., the above- Figure 1. Using 1H NMR spectroscopy, we calculated mentioned solvatochromic parameters. Therefore, re- the partial degree of substitution in each case, e.g., search is required to understand the mutual interactions DAcetate and DPropionate for the reaction with of the solvent components as well as their interaction (Ac2O + Pr2O)/IL-MS. The ratio between the indi- with cellulose. In this regard, kinetic data (rate cons- vidual third-order rate constants k3, e.g., k3,Acetate/ tants and activation parameters) help in delineating the k3,Propionate and that of the corresponding partial degree reason, e.g., for the efficiency of particular combina- of substitution, e.g., DSAcetate/DSPropionate were similar. tions of different ILs and MSs. That is, kinetic data can be successfully used to pre- We studied the kinetics of cellulose uncatalyzed- and dict the partial degrees of substitution of the produced imidazole-catalyzed acylation by carboxylic acid an- mixed esters. The reaction activation parameters indi- hydrides (acetate to hexanoic) in LiCl-DMAc, pure cated that the observed order of reactivity is due to ILs and IL-MSs [8,12-15]. Here we extend these both enthalpy and entropy. studies to cover the formation of mixed esters of

Figure 1: Molecular structures of the substances of interest in the present work; acetic-, benzoic- and propionic anhydride; N-acetyl-, N-propionyl-, N-benzoylimidazole and the ionic liquid 1-allyl-3-methylimidazolium chloride.

58 LENZINGER BERICHTE 94 (2018) 57 – 66

Experimental Section increase in solution conductivity () as a function of time (t), at a constant temperature (T). Solvents, reagents and the IL The experiment was carried out by introducing the MCC The MSs and reagents employed were purchased from solution in IL-MS into the conductivity cell; thermal Alfa Aeser or Merck and were purified according to equilibration (10 min); addition of the pre-heated so- recommended procedures [22]. Powder MCC (Avicel lution of the acylating agent in the MS; following the PH 101) was obtained from FMC BioPolymer, Philadel- change of () as a function of (t). Acid anhydrides phia. The viscosity-based degree of polymerization of were dissolved in the MS, whereas N-acylimidazoles this cellulose (DPv = 175) was determined by a standard were prepared in situ by reacting imidazole with the procedure [23], using shear-dilution Cannon-Fenske appropriate acid anhydride in the MS (molar ratio viscometer (Schott), inserted in Schott AVS 360 auto- imidazole/acid anhydride 2:1, 30 minutes stirring, room matic viscosity determination equipment. To determine temperature; clear solutions). Table SI-1 (Table 1 of the MCC index of crystallinity (Ic) we used Rigaku Supplementary Information) shows the final concen- Miniflex diffractometer, operating at 30 kV, 15 mA and trations of MCC, the IL and the acylating agents in the (Cuk) = 0.154 nm; 0.02°/minute was. Ic was calcu- conductivity cell. -1 lated using the peak height method,[24] Ic = 0.82. The observed rate constants (kobs, s ) were calculated The IL was synthesized by a microwave-assisted sol- from the dependence of () on (t), see Figure 2. The -1s-1 ventless procedure (CEM, model DU-8316), as given second order rate constant (k2; L mol ) was calculated 1 elsewhere [7]. Its purity was confirmed by H NMR from k2 = kobs/[acylating agent]. Dividing k2 by [IL] -2 -1 spectroscopy (Varian Innova-300 NMR spectrometer; gave the third order rate constants (k3, L2mol s ). The 300 MHz for 1H, reference TMS) Its solution in water activation parameters were calculated from the de- showed the absence of acid- or base-impurity (Hanna pendence of k3 on T (10 to 70 °C), by using standard H198103 portable pH-meter). equations [9].

Dissolution of cellulose in binary mixtures Synthesis of cellulose mixed esters under of ionic liquid/molecular solvents conditions of the kinetic experiment MCC (0.35 g, 21.6 mmol) was weighed into a three- The kinetic experiments were repeated on a larger scale necked round-bottom flask fitted with magnetic stirrer as described elsewhere [20]. Briefly, 0.15 g of MCC was and addition funnel (no pressure equilibration tube) dissolve in 5 mL of 4.24 mol/L IL in the appropriate containing the appropriate mass of pure IL (20.18 g; MS using microwave heating (80 °C, 30 min, 30 W), 127 mmol). The flask was connected to a vacuum followed by addition of 5 mL of 1.1 mol/L of Ac2O/ pump (2 mmHg) and heated to 90 ºC in ca. 2 h. While Pr2O (1:1 molar ratio) or AcIm/BzIm (1: 3 molar ratio) maintaining the reduced pressure, the IL was slowly in the MS and heating (80 °C, 4 h, 30 W). The produced introduced, with continuous stirring. The mixture cellulose mixed ester was precipitated in ethanol, and (MCC plus IL) was kept under these conditions for further purified by suspension several times in hot 2 h. After complete dissolution of the MCC, the appro- ethanol. It was filtered and dried for 24 h at 60 °C under priate mass of the MS (30.4 g of DMAc or 29.82 g reduced pressure. Values of the partial DS of the proof DMSO) was added and the solution stirred for 15 duct were calculated from their 1H NMR spectra (Varian minutes. The compositions of the resulting clear, Innova-300 NMR spectrometer), using DMSO-d6 to slightly amber cellulose solutions were, in mol/L: which a drop of trifluoroacetic acid was added to dis- MCC = 0.048; IL = 2.82, MS = 7.75, 8.48 for DMAc place the signal from the residual OH groups of the and DMSO, respectively. Because of miscibility AGU. The partial DS values were calculated by com- problem for the reaction with Ac2O/Pr2O in DMAc, paring the integration the methyl groups of the acetate we used a more dilute MCC solution, 0.036 mol/L. and propionate moieties, see Figure SI-1 (Figure 1 of Supplementary Information), or the acetate and ben- Kinetics of acylation of MCC in IL/MSs zoate groups, see Figure 2. For conductivity measurements, we employed PC- controlled Mettler-Toledo S400 SevenExcellence pH/mV meter provided with Mettler InLab 710 conductivity electrode. The latter was inserted into the PTFE cover (fitted tightly with aid of a Viton O-ring) of a double-walled conductivity cell through which water was circulated from a thermostat. We monitored the progress of the acylation reaction by following the

59 LENZINGER BERICHTE 94 (2018) 57 – 66

of the two esters was determined by gas chromato- graphic analysis using Shimadzu GC-17A equipment, FID, helium carrier gas, flow rate 71 mL/min, injector T = 250 °C, column T = 85 °C for 3 min and heating ramp of 20 °C/min until 250 °C.

Calculation of the volumes of solvated N-acylimidazoles Calculation of the volumes of N-acylimidazoles was done by geometry optimization (gas) using the Orca v. 2.9 software [25]. Molecular volumes of the optimized geometries of N-acylimidazoles in DMSO were computed using the Gaussian 09 software [26].

Figure 2: 1H NMR spectrum in DMSO-d6 of acetate/benzoate RESULTS AND DISCUSSION mixed ester synthetized in IL/DMSO (1 AcIm + 3 BzIm/AGU:). DSAc = 1.2; DSBz = 1.08; DSAc/DSBz = 1.11. The values DSAcetate Kinetic data: rate constants and and DSBenzoate were calculated by comparing the integrations activation parameters of CH3CO- (1.9-2.1 ppm) and C6H5CO- (7.6-8.2 ppm) peaks. We carried out the kinetic experiments under pseudo first-order conditions; the concentrations of the acylat- Synthesis of acetate/benzoate mixed esters ing agents and the IL were in excess relative to [MCC]. of cyclohexylmethanol Figure 3 shows the variation of solution conductivity This ester was synthesized by reacting mixtures of as a function of (t). Values of kobs were calculated N-acetyl- and N-benzoylimdazole in in different molar from (linear) plots of ln( - t) versus (t), where the ratios with cyclohexylmethanol (CHM, molar ratio: “infinity reading” ) was calculated by non-linear least alcohol:N-acylimidazole = 3:1) using microwave heat- square analysis. Second-order rate constants (k2) were ing (60 °C, 2 h, 30 W) in acetonitrile. The relative yields calculated by dividing (kobs) by [acylating agent].

Figure 3: Typical kinetic results showing the dependence on time (t) of solution conductance, employed to calculate the pseudo- first order rate constants (kobs). Part (A) acylation of MCC with pure AcIm (40 °C, IL/DMAc); (B) reaction of MCC with AcIm

+ BzIm (40 °C, IL/DMAc); (C) reaction of MCC with Ac2O (50 °C, IL/DMSO); (D) reaction of MCC with Ac2O + Pr2O.

60 LENZINGER BERICHTE 94 (2018) 57 – 66

Previously we showed that (k2) increases as a function of MCC with each acylating agent alone, and then with of increasing [IL] because MCC reacts as a complex a mixture of both acylating agents. The calculated rate (AGU-OH…..IL), so that we calculated the value of constants are listed in Table 1; the corresponding acti- (k3) from: k3 = k2/[IL] [8]. vation parameters are given in Table 2. The acylating agents used included acid anhydrides and N-acylimidazoles. In each case, we ran the reaction

2 -2 -1 Table 1: Calculated third order rate constants (k3; L mol s ) for the acylation of microcrystalline cellulose with carboxylic acid anhydrides in mixtures of 1-allyl-3-methylimidazolium chloride (IL) and the molecular solvents N,N-dimethylacetamide (DMAc) and dimethyl sulfoxide (DMSO).a

1 a - The ratios DSAcetate/DSPropionate, calculated from the H NMR spectra of the synthesized cellulose mixed esters were 1.1 and 1.3 for the reaction in IL-DMAc and IL-DMSO, respectively. b- The numbers within parenthesis in the fourth column correspond to % difference between k3 for the acylation by mixture of anhydrides and the mean of k3 for the reaction with the individual anhydrides. That is, the value within parenthesis =

{[k3(Ac2O+Pr2O) – [(k3Ac2O + k3Pr2O)/2)]/ k3 (Ac2O+Pr2O)} x 100.

61 LENZINGER BERICHTE 94 (2018) 57 – 66

Table 2: Calculated activation parameters for the acylation of microcrystalline cellulose with carboxylic acid anhydrides in mixtures of 1-allyl-3-methylimidazolium chloride (IL) and the molecular solvents N,N-dimethylacetamide (DMAc) and dimethyl sulfoxide (DMSO).a

a- Activation parameters were calculated at 40 °C

The corresponding kinetic data for acylation of MCC by N-acylimidazoles in IL-MS are listed in Tables 3 and 4, respectively.

Table 4: Calculated activation parameters for the acylation of microcrystalline cellulose with N-acylimidazoles in mixtures of 1-allyl-3-methylimidazolium chloride (IL) and the molecular solvents N,N-dimethylacetamide (DMAc) and dimethyl sulfoxide (DMSO).a

a- Activation parameters were calculated at 40 °C

62 LENZINGER BERICHTE 94 (2018) 57 – 66

2 -2 -1 Table 3: Calculated third order rate constants (k3 in L mol s ) for the acylation of microcrystalline cellulose with N-acylimidazoles in mixtures of 1-allyl-3-methylimidazolium chloride (IL) and the molecular solvents N,N-dimethylacetamide (DMAc) and dimethyl sulfoxide (DMSO).a

1 a - The ratio DSAcetate/DSBenzoate, calculated from H NMR spectrum (300 MHz) of the synthesized mixed esters were 0.41 (molar ration AcIm/BzIm = 1:4), 0.91 (AcIm/BzIm = 1:3) for DMAc and 0.71 (AcIm/BzIm = 1:4 BzIm), 1.11 (AcIm/BzIm = 1: 3) for DMSO. b - The numbers within parenthesis correspond to % difference between k3 for the acylation by mixture of both anhydrides and the mean of k3 for the reaction with the anhydrides separately. That is, value within parenthesis = {[k3 (AcIm+BzIm) – ((k3AcIm

+ k3 BzIm)/2)]} x 100.

63 LENZINGER BERICHTE 94 (2018) 57 – 66

Regarding these results, the following is relevant: tained using pure N-acetylimidazole and that prepared i- First we address acylation by carboxylic acid anin situ are essentially the same [14]. hydrides. In both MSs, acylation by pure Ac2O is faster than by pure Pr2O by 14  3% (DMAc) and 20  6% Regarding acylation by N-acylimidazoles (DMSO). This difference in reactivity is probably due the following is relevant: to a combination of: (a) slightly higher electrophilicity v- The (theoretically) calculated molar volumes of of the acyl group of former anhydride (pKa in water = N-acylimidazoles solvated in DMSO are 78.25, 105.37, 4.75 and 4.87, for acetic- and propionic acid, respec- 126.49 cm3/mol, giving relative volumes of 1: 1.35: 1.62 tively) [27]; (b) smaller molar volume of the former for AcIm, PrIm, BzIm, respectively. The pKa of benzoic anhydride (94.35 cm3/mol) than the latter one (128.99 acid in water is 4.2 [27], i.e., the acyl group of BzIm is 3 cm /mol), i.e., Ac2O is 73.14% smaller than Pr2O [27]. somewhat more electrophilic than that of the aliphatic ii- The observation that k3 for the reaction of MCC with N-acylimidazoles. Based on this factor alone, for AcIm/ the mixture of anhydrides is not far from the mean of BzIm of 1:1, we expected DSAcetate/DSBenzoate ≤ 1; this the sum of k3 for the reaction with the individual an- was not the case. Therefore, relative to AcIm the reac- hydrides shows that these acylations are practically in- tivity of BzIm, hence DSAcetate/DSBenzoate is determined dependent, being affected only by electronic (electro- by the balance between favorable reactivity of BzIm and philicity of the acyl group) and steric factors (volume unfavorable steric factor due to its larger volume and of the acylating reagent). rigid structure. Steric effects are important in cellulose iii- The effect of the MS is shown in the activation derivatization because cellulose dissolved in ILs is pre- parameters of Table 2. Here the reaction in IL-DMSO sent as aggregates [28,29]. Consequently, the diffusion is faster than in IL-DMAc because of gain in H of BzIm into these aggregates is expected to be slower (HDMSO < HDMAc by 5.6  0.1 kcal/mol) not than that of AcIm, leading to DSAcetate/DSBenzoate > 1, completely compensated by loss in the entropy term for AcIm/BzIm molar ratio of unity. (| TS | DMSO > | TS | DMAc by 3.8  0.1 kcal/ vi- To assess the relative importance of these opposing mol). factors we synthesized esters by reacting CHM (a model iv- The ratios between k3Ac2O/k3Pr2O in both MSs for C6-OH of the AGU) with mixtures of N-acylimida- are between 1.11 and 1.26 whereas the DSAcetate/ zoles in IL-MSs, see Table 5. In this reaction, two esters 1 DSPropionate, calculated from the H NMR spectra of are produced, namely, CHM-based (acetate + propio- the synthesized cellulose mixed esters were 1.1 and nate) and/or (acetate + benzoate). Because CHM is mo- 1.3, for DMAc and DMSO, respectively. That is, kine- nomeric we expected, in principle, CHM-acetate/ tic data can be employed to predict the structure of the CHM-second ester ≈ 1. product. To our knowledge, this is the first time that vii- Entries 1 and 2 of Table 5 indicate that the ratios of this use is shown, at least for derivatization under the esters prepared in IL-MSs are close to unity (1.1, homogeneous reaction conditions. 1.3 for DMAc and DMSO, respectively). That is, the To expand the scope of our investigation, we studied the nature of the acylating agent (aliphatic anhydride or acylation by N-acylimidazoles because these are more N-acylimidazole with aliphatic acyl group) apparently reactive than the corresponding acid anhydride and, does not affect the relative ratios between the esters. unlike acyl chlorides, no HCl is produced. N-acylimi- Entries 3 to 8 show, however, that this is not the case for dazoles were generated in situ, as shown in Scheme 1: the reaction of CHM with (AcIm + BzIm), where the former is clearly favored. Thus, the above-mentioned steric retarding effect is also operative for a simple molecule like CHM; ester ratio ≈ 1 is obtained only when BzIm is used in excess, e.g., entry 5. viii- As expected, the reaction with N-acylimidazoles is Scheme 1: Generation in situ of N-acylimidazole by the reac- faster than that with the corresponding anhydride. As tion of the diazole with acid anhydride. shown in Tables 2 and 4 the reason is a large decrease in reaction enthalpy (ca. 10 kcal/mol) not completely As shown in Scheme 1, the acylation solution contains compensated by the TS term (ca. 8 kcal/mol). The imidazolium carboxylate side product that may, in prin- reaction rates in the molar polar DMSO are 14 to 29% ciple, catalyze the acylation reaction by acid-base cataly- faster than in DMAc due to subtle differences in the sis. Previously we showed, however, that imidazolium activation parameters. acetate is inert in acylation, i.e., the rate constants ob-

64 LENZINGER BERICHTE 94 (2018) 57 – 66

Table 5: Conditions for- and results of the microwave-assisted synthesis of cylohexylmethanol-based esters by reaction with N-acylimidazoles in IL-MSs.a

a - Ester ratio refers to the ratio between the two esters produced simultaneously (always relative to acetate). E.g., for the reaction of CHM with (AcIm + PrIm) the result of entry 1, column 4 means that the yield cyclohexylmethyl acetate/ cyclohexylmethyl propionate = 1.1.

group. The use of kinetic data to predict product com- Conclusions position eliminates the labor involved optimizing the reaction conditions to obtain mixed esters with targeted Using a simple technique, conductivity, we studied the partial DS values. formation of mixed esters, namely acetate/propionate and acetate/benzoate obtained by the reaction of MCC Acknowledgements with mixtures of acid anhydrides and/or N-acylimidazoles. This is the first report on the kinetics of forma- Thais Bioni thanks CAPES pre-doctoral research tion of mixed esters under homogeneous reaction con- fellowship, O. A. El Seoud thanks CNPq and FAPESP ditions. The composition of the produced esters can for research productivity fellowships (grant 307022/ be predicted from the kinetic data, if steric factors are 2014-5), and financial support (grant 2014/22136-4). not dominating. This is the case for flexible acylating Naved I. Malik thanks FAPESP for a postdoctoral agents, e.g., Ac2O and Pr2O. The reactivity of rigid research fellowship (grant 2013/18008). We thank acylating agents, e.g., BzIm is controlled by steric Dr. Paulo A. R. Pires for carrying out the theoretical effects more than by the electrophilicity of its acyl calculations.

65 LENZINGER BERICHTE 94 (2018) 57 – 66

References

[1] M. Gericke, P. Fardim, T. Heinze, Molecules, [16] P. Carrollo, B. Grospietro, Plastic Materials Macro- 2012, 17, 7458. mol. Symp., 2004, 208, 335. [2] M. Kostag, K. Jedvert, C. Achtel, T. Heinze, O. A. [17] K. J. Edgar, C. M. Buchanan, J. S. Debenham, El Seoud, Molecules, 2018, 23, 511. P. A. Rundquist, B. D. Seiler,M. C. Shelton, [3] T. Heinze, O. A. El Seoud, A. Koschella, Cellulose D. Tindall, Prog. Polym. Sci., 2001, 26, 1605. derivatives: Synthesis, structure, and properties, [18] R. T. Bogan, R. J. Brewer, Cellulose esters, organic. 1a Ed., Springer, Cham, 2018. In: J. I. Kroschwitz, M. Bickford, A. Klingsberg, [4] K. Baba, E. Ono, S. Itoh, K. Utoh, T. Noda, K. Usui, J. Muldoon, A. Salvatore, (eds) Encyclopedia of M. Ishihara, H. Inamo, D. Takagi, T. Asano, J. polymer science and engineering. Wiley, NY, 1985, Chem. Eur., 2006, 12, 5328. 158. [5] S. Tiwari,A. Kumar, J. Phys. Chem. A, 2012, 116, [19] L. C. Treece, G. I. Johnson, Chem. Ind., 1993, 49, 1191. 241. [6] N. D. Khupse, A. Kumar, J. Phys. Chem. A, 2011, [20] S. Possidonio, L. C. Fidale, O. A. El Seoud, 115, 10211. J. Poly. Sci. Part A: Poly. Chem., 2010, 48, 134. [7] T. A. Bioni, E. P. G. Arêas, L. G. Couto, G. Favarin, [21] L. C. Fidale, S. Possidonio, O. A. El Seoud, Macro- O. A. El Seoud, Nordic Pulp, 2015, 30 (1), 105. mol. Biosci., 2009, 9, 813. [8] H. Nawaz, P. A. R. Pires, E. P. G. Arêas, N. I. Malek, [22] W. L. F. Armagero, C. L. L. Chai, Purification of O. A. El Seoud, Macromol. Chem. Phys., 2015, Laboratory Chemicals, 5th ed., Elsevier Ltd, New 216, 2368. York, 2003. [9] E. V. Anslyn, D. A. Dougherty,Modern Physical [23] ASTM, D 1795-94, Standard test methods for in- Organic Chemistry, University Science Books, trinsic viscosity of cellulose, 2001. Sausalito, 2006, 355. [24] G. Buschlediller, S. H. Zeronian, J. Appl. Polym. [10] N. D. Khupse, A. Kumar, J. Solution Chem., Sci., 1992, 45, 967. 2009, 38, 589. [25] F. Neese, U. Becker, D. Ganiouchine, S. Koβmann, [11] M. Tariq, K. Shimizu, J. M. S. S. Esperança, J. N. T. Petrenko, C. Riplinger, F. Wennmohs, Orca C. Lopes, L. P. N. Rebelo, Phys. Chem. Chem. package v, 2.9, University of Bonn, 2011. Phys., 2015, 17, 13480. [26] Gaussian, Inc., Wallingford CT, 2016. [12] P. A. R. Pires, N. I. Malek, T. C. Teixeira, T. A. [27] CRC Handbook of Chemistry and Physics, ed. D. Bioni, H. Nawaz, O. A. El Seoud, Industrial Crops R. Lide, CRC Press, Boca Raton, USA, 85th edn, and Products, 2015, 77, 180. 2004. [13]H. Nawaz, P. A. R. Pires, T. A. Bioni, E. P. G. [28] P. C. Trulove, W. M. Reichert, H. C. De Long, S. R. Arêas, O. A. El Seoud, Cellulose, 2014, 21, 1193. Kline, S. S. Rahatekar, J. W. Gilman, M. Muthu- [14] H. Nawaz, P. A. R. Pires, O. A. El Seoud, Carb. kumar, ECS Trans., 2009, 16, 111. Poly., 2013, 92, 997. [29] O. Kuzmina, E. Sashina, S. Troshenkowa, [15] H. Nawaz, R. Casarano, O. A. El Seoud, Cellulose, D. Wawro, Fibers Text East Eur., 2010, 18, 32. 2012, 19, 199.

66 LENZINGER BERICHTE 94 (2018) 67 – 76

Lignocellulosic Multicomponent Fibers Spun from Superbase-Based Ionic Liquids

Michael Hummel*, Yibo Ma, Anne Michud, Shirin Asaadi, Annariikka Roselli, Agnes Stepan, Sanna Hellstén, and Herbert Sixta*

Department of Bioproducts and Biosystems, Aalto University, P.O. Box 16300, FI-00076 Aalto, Finland *Contact: [email protected], [email protected]

In memoriam Mag. Dr. Gerhard Laus († 2017) Abstract

Recently, the list of solvents for cellulose dissolution and fiber spinning was extended with so-called ionic liquids, which bear great potential in terms of environmental and technical aspects. Herein, we report on 1,5-diazabicyclo[4.3.0]non-5-ene-1-ium acetate – a non-imidazolium, super-base-based ionic liquid – as excellent solvent for a wide set of lignocellulosic solutes to prepare composite fibers of cellulose, hemicellulose, and lignin. The solutions are processed in a Lyocell-type dry-jet wet spinning procedure, which allows for a filament draw in the air gap. The effect of the draw was dependent on the solute composition. With a high share of cellulose, a small draw led already to high tensile strength and modulus. All fibers showed high mechanical properties even when containing a high share of non-cellulosics.

Keywords: lignocellulose, multicomponent fibers, ionic liquids, spinning, fibers

Introduction

In 2005, Lenzinger Berichte published the first report cellulose-dissolving ILs, some of those were also on man-made cellulosic fibers spun from ionic liquids identified as suitable solvents for the secondary wood (ILs). [1] This seminal study by Laus and Schotten- polymers, hemicellulose and lignin, either as isolated berger was motivated by an earlier publication of the polymers or jointly in form of the entire wood matrix. Rogers group in 2002, when they reported the disso- Extensive reports and reviews on cellulose and wood lution of cellulose in imidazolium-based ionic liquids dissolving ILs are meanwhile available. [5-10] and postulated shaping processes such as filament Following the aforementioned report of Laus et al. in spinning. [2] ILs consist entirely of ions with a defined Lenzinger Berichte, numerous studies have investi- melting point below 100 °C. Through the combina- gated spinning of different cellulosic solutes in various, tion of sterically demanding organic cations (mostly almost exclusively imidazolium-based ILs. A detailed with a delocalized charge) and organic or inorganic review was given earlier. [4] In two extensive studies, anions the fluid state is favored in terms of enthalpy the miscibility and processability of polysaccharide- and entropy, giving those salts their characteristic low polysaccharide mixtures in [emim]OAc was compared melting points. Amongst the virtually infinite number to NMMO and caustic solution. [11-12] Natural poly- of ILs accessible through the combination of cations mers were found to have a strong affinity to cellulose and and anions [3], only few have been identified as cel- gave homogeneous multicomponent fibers. Contrary, lulose solvents. Most of them are still based on an only few reports address lignocellulosic composite imidazolium cation and have anions such as halides, fibers. [13] Fink and coworkers have produced con- carboxylates, or phosphates to name only the most tinuous filaments by dissolving hemp entirely [14] prominent ones. [4] Shortly after the discovery of and have shown that cellulose can be co-processed with

67 LENZINGER BERICHTE 94 (2018) 67 – 76

lignin added to the cellulose solution. [15] fibers. The archetypical IL, [emim]OAc, has been The vast majority of the studies describing IL-based promoted as excellent cellulose solvent due to its low fiber spinning are based on imidazolium halide or viscosity. However, filaments extruded from [emim] carboxylate ILs and used rather simple extrusion equip- OAc-cellulose solutions can only be stretched to a ment which yielded filaments with diameters of 100 limited extent because they lack the required elasticity. µm or higher. Unfortunately, they bear little potential Also, the coagulation mechanism has to favor the ra- to advance from proof-of-concept extrusion trials to pid development of a solid polymer filament to pre- industrial spinning processes as several limitations and vent breaches in the spin bath. [27] Choosing wet drawbacks have been identified during the past decade. spinning for IL-biopolymer solutions allows to use Already in the initial Lenzinger Berichte [1,16], severe smaller spin capillaries with smaller diameter as sub- polymer degradation was observed, when cellulose sequent draw forces are noticeably smaller, prevent- was processed in imidazolium-based ILs. Evidently, ing aforementioned filament breaks. However, the lack thermal stability and inertness towards the solute at of draw also limits the mechanical properties of the the required processing temperatures are indispensable resulting fibers. Polymer alignment is mostly confined for feasible fiber spinning. Excessive degradation of to shear-induced orientation in the spin capillary, which the IL and biopolymer does not only limit the economy is only moderate compared to draw-induced alignment. of the solvent but also complicates the solvent recovery Stretching of the filament in the spin or washing bath and recycling substantially. Mainly for this reason, as it is done for advanced viscose fibers such as tire halide-containing ILs seem to have little potential for cord or to a smaller extent for Modal fibers is also upscaling. They require comparatively high dissolution only possible within limits. The chemical re-generation and processing temperatures at which the IL decom- of viscose-type can be retarded through different poses [17-18] and cellulose is depolymerized severely additives to the dope or spin bath. The physical coagu- even if stabilizers are added. [1,16,19] Carboxylate- lation of cellulose-IL solution, however, is assumed to based ILs could bypass this problem to some extent. proceed via spinodal decomposition [28] which can- [20] However, the imidazolium moiety can initiate not be retarded to such an extent as in the viscose pro- side reactions leading to cellulose modification [21] cess. Thus, dry-jet wet spinning is the more promising such as depolymerization [22-23] or acetylation through technique for processing of ionic liquid-biopolymer trans-acylation.[24] solutions and, consequently, suitable visco-elastic Thermal and chemical stability represent only one cri- properties are mandatory. teria for an industrially applicable spinning process. The resulting ionic liquid-polymer solutions must show distinct visco-elastic properties to be suitable for fiber spinning. Many proof-of-concept studies demonstrate the extrusion of filaments with a large diameter through syringe-type devices but do not address the complexity of reducing the diameter to a level of <20 µm which is needed for textile or technical applications. A key- feature of dry-jet wet spinning (cf. Lyocell process) is the draw of the fluid filament in the air gap. The generated elongational stress on the polymer solutions causes the cellulose chains to align along the fiber axis. The highly Figure 1: Chemical structure of 1,5-diazabicyclo[4.3.0] oriented structure is preserved upon coagulation, re- non-5-ene-1-ium acetate ([DBNH]OAc) used as solvent in sulting in filaments with high mechanical strength and this study. modulus characteristic for Lyocell-type fibers. The filament draw, however, also leads to a substantial In 2013, we found that the superbase-based ionic decrease of the fiber diameter and is essential as the liquid 1,5-diazabicyclo[4.3.0]non-5-ene-1-ium aceta- diameter of the spinneret orifices can only be reduced te [DBNH]OAc (Figure 1) is an excellent solvent for to a certain extent. A set of complex breaching mecha- lignocellulosic material. [29] Although solid at room nisms that can be observed in the air gap set a lower temperature, [DBNH]OAc shows a very low viscosity limit to the initial filament diameter. [25-26] If the of ca. 23 mPa s at its melting point of 65 °C. This filament is too thin at its exit-point it will break in the facilitates handling and favors the kinetics of polymer air gap and impede spinnability. Thus, the drawability dissolution. The resulting cellulose solutions are of an initially thick filament is a key feature in the dry- characterized by visco-elastic properties that allow jet wet spinning process to yield thin and strong for dry-jet wet spinning at moderate temperatures of

68 LENZINGER BERICHTE 94 (2018) 67 – 76

70-85 °C with very high draw applicable, resulting in grammage were used as fine paper. Cuttings residues fibers with excellent mechanical properties. [DBNH] from a Finnish fluting board mill were used as cardboard OAc proved to be a robust solvent to spin a broad source. Newsprint (Helsingin Sanomat) was collected variety of lignocellulosics. Besides various dissolving from a local recycling station in Espoo, Finland. Birch- pulps [30], cellulose in combination with hemicellu- wood (Betula pendula) chips were provided by Metla, lose and/or lignin such as paper-grade kraft pulp, Finland. Beechwood organosolv lignin (OS) was grate- mildly refined birch wood [31], waste paper and card- fully received from Fraunhofer Institutes, Germany. [34] board [32], newsprint [33], and cellulose-lignin mixtures [34] were successfully spun into high-perfor- Details of the respective refining procedure of the mance fibers as reported earlier. Herein, we provide a birch wood [31], waste paper and cardboard [32], and comprehensive summary of those fibers and establish newsprint [33] were described earlier. If necessary, a relationship between the non-cellulosic part and the DP was adjusted through sulphuric acid treatment their influence on the properties of the resulting fibers. or E-beam irradiation. [DBNH]OAc was prepared from 1,5-diazabicyclo [4.3.0]non-5-ene (99%, Fluorochem, UK) and acetic Materials and Methods acid (glacial, 100%, Merck, Germany) as reported previously. [4] Materials Eucalyptus (Eucalyptus urugrandis) prehydrolysis Pulp dissolution and spinning kraft pulp ([η] = 468 ml/g, DP = 1026, Mn = 79.8 kDa, [DBNH]OAc was liquefied at 70 °C before the (lig- Mw = 268.6 kDa, polydispersity 3.4, Bahia Speciality no-) cellulosic solute was dispersed in the IL. The Cellulose, Brazil), birch (Betula pendula) prehydroly- mixture was transferred to a vertical kneader system as sis kraft pulp ([η] = 476 ml/g, DP = 1133, Mn = 65.9 described previously [10] and typically dissolved in kDa, Mw = 269.3 kDa, polydispersity 4.1, Enocell 90 min or less. Multi-filaments were spun on a custo- Speciality Cellulose, Finland), birch ECF kraft pulp mized laboratory piston spinning system (Fourné Poly- ([η]: 1025 ml/g; DP = 3100, Mn = 53.3 kg/mol, Mw = mertechnik, Germany) as described earlier (Figure 2). 625.5, polydispersity 11.7, Stora Enso – Imatra mill), [4] The spin solutions were extruded at 60-75 °C and cotton wastes in form of hospital bed sheets (452 through spinnerets of 36 or 200 holes with a capillary ml/g, DP = 1053, Mn = 151.6 kg/mol, Mw = 348.4 kg/ diameter of 100 μm and a length to diameter ratio mol polydispersity 2.29, Uusimaa Hospital Laundry (L/D) of 0.2. The filaments were drawn in a ca. 1 cm – Uudenmaan sairaalapesula Oy, Finland) were used air gap, coagulated in a water bath (15 °C), washed in as cellulosic solute. A4 copy paper sheets of 80 g/m2 hot water (60 °C), and air dried.

Figure 2: Left: schematic depiction of the customized laboratory piston spinning system (ve = extrusion velocity, vtu = take-up velocity); right: liquid filament in the air gap. After exiting the spinneret capillary the polymer solution shows dye swelling (exaggerated depiction) before the filament diameter decreases rapidly due to the draw force exerted by the increased take-up velocity. The draw ratio is defined as DR= tuv / ve. [4, 32]

69 LENZINGER BERICHTE 94 (2018) 67 – 76

Analytical methods addition of organosolv (OS) lignin to eucalyptus PHK Methods for the assessment of the rheological proper- pulp. [34] Feedstock with both hemicellulose and ties, chemical composition and molar mass distribution lignin was included via paper and cardboard samples are reported elsewhere. [31] Linear density (dtex), that were refined stepwise to increase the cellulose tenacity (cN/tex) and elongation at break (%) were content gradually. [32] Also, waste newsprint pre- determined according to DIN 53816 in both con- treated with an alkaline glycerol pulping process [33] ditioned (23°C, 50% RH) and wet states using a Vibro- and unbleached birch kraft pulps [31] served as mixed- skop-Vibrodyn system (Lenzing Instruments GmbH polymer solute. Both feedstock were pretreated with & Co KG, Austria) as described earlier. [33] The varying intensity to tailor the non-cellulosic polymer Young’s modulus of the spun fibers was calculated content. from the slope of the entire elastic region of the stress- All obtained pulps were adjusted to a target intrinsic strain curves with a Matlab script according to ASTM viscosity of 400-500 ml/g either by sulfuric acid standard D2256/D2256M. treatment or E-beam irradiation. The latter did not alter the composition of the feedstock. This particular vis- Results and Discussion cosity level is typical for pulps used in the NMMO- based Lyocell process and proved also suitable for Lignocellulosic solutes Ioncell spinning. However, the adjustment to the target In several earlier reports we have presented different intrinsic viscosity does not guarantee spinnability. In sets of fibers spun from cellulose with or without a systematic study, pulps with varying molar mass hemicellulose and lignin.[30-35] An overview of the distribution were prepared and their spinnability cor- various raw materials and their composition is given related to the macromolecular features of the respective in Table 1. Cotton taken from waste hospital bed sheets solutes.[36] It was found, that a share of 20-30% of consisted of pure cellulose and had practically no cellulose chains with a DP≥2000 and an intermediate other non-cellulosic constituents. [35] Two hardwood polydispersity index of around 3-4 were necessary to prehydrolysis kraft (PHK) dissolving pulps consisted provide the visco-elastic properties that allow for a of ca. 95% cellulose. [30] The hardwood kraft pulp high draw ratio in the air gap. Accordingly, the pulps sample had ca. 25% hemicelluloses. Pulps with a described in Table 1 were adjusted to meet those re- lignin content of 10-50% were prepared through quirements.

Table 1: Composition of raw material and fibers spun from [DBNH]OAc.

70 LENZINGER BERICHTE 94 (2018) 67 – 76

Fiber spinning nuous extrusion and potential to draw the fluid fila- After proper tuning of the rheological properties ment in the air gap. [37] Herein, a draw ratio of >10 through polymer DP adjustment, variation of solute was considered as good spinnability, which was exhi- concentration and selection of the spin temperature all bited by all samples (see Table 2). All solutions could presented solutes showed good spinnability. As men- be spun at 60-75 °C. tioned earlier, spinnability is defined by stable conti-

Table 2: Properties of fibers spun from 13% solutions of the various solutes at the indicated draw ratio (spin capillary diameter in all cases: 100 µm). The pulp code is given in Table 1.

Due to the mild processing conditions and high inertness of the ionic liquid the lignocellulosic solutes did not undergo significant depolymerization as reported for other ILs. [1,16,20] The molar mass distributions of the holocellulosic constituent were virtually identical for the original solute and the spun fiber. [4, 32] Compositional analysis revealed that a very small share of hemicelluloses was not recovered in the spun fiber and lost in the spin bath. It was shown earlier that low molecular weight hemicelluloses can be solubilized by IL-water mixtures. [38-39] Also, in the case of fibers with added beech lignin, lignin was not entirely recovered and a small amount was detected in the spin bath. Consequently, the cellulose content of the fibers is slightly higher than in the original solute (Table 1). Figure 3: Stress-strain curves of selected samples produced Stress-strain curves of some selected samples and of at highest DR. For comparison, commercial Lyocell and commercial Lyocell and regular viscose fibers are plot- viscose fibers were included. ted in Figure 3. Viscose as the only wet spun fiber plotted in Figure 3 shows the lowest slope in the initial 8-10% typical for Lyocell-type fibers but higher elastic part, the highest elongation, and lowest tenacity. tenacity than the commercial NMMO-Lyocell fiber. All Ioncell fibers show a similar elongation between At 50% organosolv lignin content, the fibers show a

71 LENZINGER BERICHTE 94 (2018) 67 – 76

reduced tenacity, which is still higher than that of a value which depends on the solute composition. In commercial viscose fiber. At the same time, the elonga- Figure 4, the lignocellulosic materials are roughly tion was extended to about 16%. grouped according to their cellulose content. With lignin and hemicellulose being comparatively short Fiber properties biopolymers, the tensile properties are governed by As the draw ratio is getting higher, the lateral orienta- the cellulosic constituent. With increasing cellulose tion of the polymer chains increases. This has a pro- content higher fiber tenacities are accessible with cot- nounced effect on the mechanical properties of the re- ton as 100% cellulose and the 50/50 cellulose-lignin sulting fibers. Figure 4 illustrates the development of the fibers representing the upper and lower limits, respec- tenacity as a function of the draw ratio. All solutions tively. This relationship between the cellulose content could be drawn to a high extent, virtually independent and the tensile strength is known and used for com- of the cellulose content. As the draw is getting higher, mercial technical fibers such as tire cords, which are the tenacity starts to level off approaching a plateau derived from high purity dissolving pulps.

Figure 4: Development of the conditioned tenacity with increasing draw ratio. The lignocellulosic solutes are grouped according to their cellulose content (green, closed symbols: ca. 95% cellulose; blue, open symbols: ca. 90%; orange, +centered symbols: ca. 75-80%; red, half-filled symbols: ca. 50-70%).

Remarkably, the draw required to approach the The Young’s moduli of the spun fibers – deduced from plateau value depends on the share of non-cellulosic the initial elastic part of the stress-strain curve – show a constituents. Filaments based on cotton, i.e. 100% similar trend as observed for the tenacity. Evidently, as cellulose, develop tenacity very quickly and already the share of cellulose polymer chains and their orienta- at a moderate draw of 5.3 they reach 90% of their final tion increase the resulting matrix is gaining in rigidity, tensile strength of 58 cN/tex (at DR 14.1). As the cel- reflected by an elevated modulus. Non-cellulosic poly- lulose content decreases, the tenacity-vs-draw curves mers act as softener and reduce the tensile modulus flatten out and the tenacity is only developing slowly. of the composite fibers. For pure cellulose fibers the A similar behavior was observed in a previous study increase of the modulus can be correlated with an in- when varying the concentration of dissolving pulp. crease in the tensile strength. [40] A similar trend is The tenacity developed much faster when a 17% solu- observed to some extent despite the composite nature tion was spun as compared to 10% cellulose concent- of the fibers under investigation (Figure 5). Based on ration, where the fibers gained strength only slowly previous interpretations, this relationship might indi- with increasing draw ratio. [29] cate high structural conformity of the multi-component

72 LENZINGER BERICHTE 94 (2018) 67 – 76

fibers, that is the polymers are evenly distributed within some deviation is observed for fibers with a higher the fiber forming a homogeneous matrix. However, share of hemicelluloses.

Figure 5: Relationship between tenacity and Young’s modulus (in conditioned state) of all presented fibers at different draw ratios.

The elongation at break of regenerated cellulosic fibers is an important parameter in particular when targeting textile applications. The stiffness of the fibers is affecting the resulting yarn and fabrics and might lead to a reduced softness feel of the final gar- ment. This is sometimes perceived as a downside of Lyocell-type filaments and other fibers with a highly oriented structure which are characterized by an elongation at break of <15% in conditioned state. [41] Lenz at al. proposed that the elongation mechanism during lateral fiber deformation is associated with the orientation of the crystalline regions and the straight- ening of the less ordered amorphous domains in the interstices between the crystallites. [42-44] Thus, the draw ratio also affects the elongation at break. In con- Figure 6: Dependency of the conditioned tenacity of fibers trast to the above discussed tensile strength and modu- spun at draw ratio of 10 on the cellulose content of the lus, the elongation declines for all lignocellulosic different materials. The inset shows the correlation of the solutes as the draw ratio is increased. Non-cellulosic tenacity with the share of low-molecular weight cellulose constituents can act as plasticizers and increase the chains with a DP<100. elongation at break. Figure 6 illustrates the tensile strength of all fibers as Summary of the influence of non-cellulosic a function of their cellulose content. In order to mini- constituents on the fiber properties mize draw-induced differences in the cellulose orien- As described above, cellulose is the property-determin- tation, all depicted values derive from fibers spun at ing constituent in the polymer blend fibers. Owing to DR 10. The orientation of fibers with non-cellulosic its comparably high DP, cellulose governs in particular constituents will be reduced nonetheless. As mentioned the mechanical properties. A strength weakening effect earlier, the mechanical properties are connected to of an increasing share of hemicellulose was also re- various structural parameters such as crystallinity, ported for viscose and NMMO-Lyocell fibers. [45-46] crystallite length and orientation of the amorphous

73 LENZINGER BERICHTE 94 (2018) 67 – 76

interstice-domains. [44] Although they are primarily References affected by the draw, non-cellulosic parts will obviously influence them as well. [1] Laus, G.; Bentivoglio, G.; Schottenberger, H.; A distinct correlation can be observed, with cotton as Kahlenberg, V.; Kopacka, H.; Roeder, H.; Roeder, pure cellulose yielding the highest strength value. In T.; Sixta, H., Ionic liquids: current developments, fact, it is not the chemical nature of hemicellulose that potential and drawbacks for industrial applica- affects the fiber properties but the share of short-chain tions. Lenzinger Ber. 2005, 84, 71-85. polymers which is proportional to the presence of [2] Swatloski, R. P.; Spear, S. K.; Holbrey, J. D.; hemicellulose. [47] The inset in Figure 6 shows the Rogers, R. D., Dissolution of Cellose [sic] with tenacity of the holocellulose fibers vs. the share of Ionic Liquids. J. Am. Chem. Soc. 2002, 124 (18), polymers with a DP<100 (cellulose-lignin blend fibers 4974-4975. are excluded). As the relative amount of short chains [3] Welton, T., Room-Temperature Ionic Liquids. is increasing the fiber tenacity is reduced. This would Solvents for Synthesis and Catalysis. Chem. suggest that narrowly distributed pulps with a high Rev. 1999, 99 (8), 2071-2083. average DP would be most suitable for the production [4] Hummel, M.; Michud, A.; Tanttu, M.; Asaadi, S.; of high-strength fibers. However, this is partly in con- Ma, Y.; Hauru, L. J.; Parviainen, A.; King, A. T.; flict with the rheological requirements where pulps Kilpeläinen, I.; Sixta, H., Ionic Liquids for the Pro- with a broad molecular weight distribution yield vis- duction of Man-Made Cellulosic Fibers: Opportuni- co-elastic properties that provide better spinnability. ties and Challenges. Adv. Polym. Sci. 2015, 1-36. [5] Pinkert, A.; Marsh, K. N.; Pang, S.; Staiger, M. Conclusions P., Ionic Liquids and Their Interaction with Cel- lulose. Chem. Rev. 2009, 109 (12), 6712-6728. More than 10 years after the seminal publication by [6] Maeki-Arvela, P.; Anugwom, I.; Virtanen, P.; Laus and Schottenberger on the potential of ionic Sjoeholm, R.; Mikkola, J. P., Dissolution of ligno- liquids for the production of man-made cellulosic fibers, cellulosic materials and its constituents using ionic superbase-based ILs seem promising to overcome the liquids-A review. Ind. Crops Prod. 2010, 32, 175-201. obstacles that were identified initially. [DBNH]OAc [7] Wang, H.; Gurau, G.; Rogers, R. D., Ionic liquid proved as excellent lignocellulose solvent to produce processing of cellulose. Chem. Soc. Rev. 2012, biopolymer composite fibers for textile and technical 41 (4), 1519-1537. applications. The wood constituents cellulose, hemi- [8] Brandt, A.; Grasvik, J.; Hallett, J. P.; Welton, T., cellulose, and lignin are well soluble and fibers with Deconstruction of lignocellulosic biomass with high mechanical properties could be spun in all cases. ionic liquids. Green Chem. 2013, 15 (3), 550-583. A distinct relationship between the cellulose content [9] Singh, S.; Simmons, B. A. In Ionic liquid pre- and tenacity, modulus and elongation at break was treatment: mechanism, performance, and challen- identified. Meanwhile, also other groups have picked ges, John Wiley & Sons Ltd.: 2013; pp 223-238. up and expanded the portfolio of super-base based ILs [10] Hauru, L. K. J.; Ma, Y.; Hummel, M.; Alekhina, for fiber spinning. [48] However, more systematic stu- M.; King, A. W. T.; Kilpelaeinen, I.; Penttilae, P. dies are certainly needed to obtain a more conclusive A.; Serimaa, R.; Sixta, H., Enhancement of ionic picture on the fiber structure formation. A detailed liquid-aided fractionation of birchwood. Part 1: study of the amorphous and crystalline orientation, in- autohydrolysis pre-treatment. RSC Adv. 2013, 3, cluding the effects of crystallinity and crystallite di- 16365-16373. mensions in the context of composite fibers is needed to [11] Wendler, F.; Meister, F.; Wawro, D.; Wesolowska, provide a better understanding of the factors governing E.; Ciechanska, D.; Saake, B.; Puls, J.; Le Moigne, the mechanical properties. N.; Navard, P., Polysaccharide blend fibres formed Further, the IL recycling concept has to be customized from NaOH, N-methylmorpholine-N-oxide and when using multiple solutes. Currently, major efforts are 1-ethyl-3-methylimidazolium acetate. Fibres Text. devoted to develop a scalable strategy for a virtually East. Eur. 2010, 18, 21-30. quantitative recovery of the IL for cellulose processing. [12] Wendler, F.; Persin, Z.; Stana-Kleinschek, K.; When using low-molecular weight polymers such as Reischl, M.; Ribitsch, V.; Bohn, A.; Fink, H.-P.; hemicellulose and lignin small shares are lost in the spin Meister, F., Morphology of polysaccharide blend bath and have to be removed to avoid accumulation over fibers shaped from NaOH, N-methylmorpholine- time. Yet, the unique opportunities offered by ILs should N-oxide and 1-ethyl-3-methylimidazolium acetate. certainly motivate to find also solutions for this challenge. Cellulose 2011, 18, 1165-1178.

74 LENZINGER BERICHTE 94 (2018) 67 – 76

[13] Sun, N.; Li, W.; Stoner, B.; Jiang, X.; Lu, X.; Ro- [25] Ziabicki, A.; Takserman-Krozer, R., Mechanism gers, R. D., Composite fibers spun directly from of breakage of liquid threads. Kolloid-Z.u.Z. solutions of raw lignocellulosic biomass dis- Polymere 1964, 198 (1-2), 60-65. solved in ionic liquids. Green Chem. 2011, 13, [26] Ziabicki, A.; Takserman-Krozer, R., Effect of rheo- 1158-1161. logical factors on the length of liquid threads. [14] Lehmann, A.; Bohrisch, J.; Protz, R.; Fink, H.-P. Kolloid-Z.u.Z.Polymere 1964, 199 (1), 9-13. Method for preparing lignocellulose spinning so- [27] Hauru, L. K. J.; Hummel, M.; Nieminen, K.; lution and spin regenerated fibers from it with- Michud, A.; Sixta, H., Cellulose regeneration out any initial preatreatments. WO2013144082A1, and spinnability from ionic liquids. Soft Mat. 2013. 2016, 12 (5), 1487-1495. [15] Lehmann, A.; Ebeling, H.; Fink, H.-P. Method [28] Biganska, O.; Navard, P., Morphology of cellulose for economical production of lignin-containing objects regenerated from cellulose-N-methyl- precursor fibers for use in further production of morpholine N-oxide-water solutions. Cellulose carbon fibers. WO2012156441A1, 2012. 2008, 16 (2), 179. [16] Bentivoglio, G.; Roeder, T.; Fasching, M.; Buch- [29] Sixta, H.; Michud, A.; Hauru, L. K. J.; Asaadi, S.; berger, M.; Schottenberger, H.; Sixta, H., Cellu- Ma, Y.; King, A. W. T.; Kilpeläinen, I.; Hummel, lose processing with chloride-based ionic liquids. M., Ioncell-F: A High-strength regenerated cellulose Lenzinger Ber. 2006, 86, 154-161. fibre. Nord. Pulp Pap. Res. J. 2015, 30 (1), 43-57. [17] Kosmulski, M.; Gustafsson, J.; Rosenholm, J. [30] Michud, A.; Tanttu, M.; Asaadi, S.; Ma, Y.; Netti, B., Thermal stability of low temperature ionic E.; Kääriäinen, P.; Persson, A.; Berntsson, A.; liquids revisited. Thermochim. Acta 2004, 412 Hummel, M.; Sixta, H., Ioncell-F: ionic liquid- (1-2), 47-53. based cellulosic textile fibres as alternative to [18] Meine, N.; Benedito, F.; Rinaldi, R., Thermal viscose and Lyocell. Text. Res. J. 2016, 86 (5), stability of ionic liquids assessed by potentiome- 543-552. tric titration. Green Chem. 2010, 12, 1711-1714. [31] Ma, Y.; Stubb, J.; Kontro, I.; Nieminen, K.; [19] Gazit, O. M.; Katz, A., Dialkylimidazolium Ionic Hummel, M.; Sixta, H., Filament spinning of un- Liquids Hydrolyze Cellulose Under Mild Con- bleached birch kraft pulps: Effect of pulping in- ditions. ChemSusChem 2012, 5 (8), 1542-1548. tensity on the processability and the fiber proper- [20] Michud, A.; Hummel, M.; Haward, S.; Sixta, H., ties. Carbohydr. Polym. 2018, 179, 145-151. Monitoring of cellulose depolymerization in [32] Ma, Y.; Hummel, M.; Maattanen, M.; Sarkilahti, 1-ethyl-3-methylimidazolium acetate by shear A.; Harlin, A.; Sixta, H., Upcycling of waste pa- and elongational rheology. Carbohydr. Polym. per and cardboard to textiles. Green Chem. 2016, 2015, 117 (0), 355-363. 18, 858-866. [21] Ebner, G.; Schiehser, S.; Potthast, A.; Rosenau, T., [33] Ma, Y.; Hummel, M.; Kontro, I.; Sixta, H., High Side reaction of cellulose with common 1-alkyl- performance man-made cellulosic fibres from re- 3-methylimidazolium-based ionic liquids. Tetra- cycled newsprint. Green Chem. 2018, 20 (1), hedron Lett. 2008, 49 (51), 7322-7324. 160-169. [22] Dorn, S.; Wendler, F.; Meister, F.; Heinze, T., Inter- [34] Ma, Y.; Asaadi, S.; Johansson, L.-S.; Ahvenainen, actions of ionic liquids with polysaccharides - 7: P.; Reza, M.; Alekhina, M.; Rautkari, L.; Michud, Thermal stability of cellulose in ionic liquids and A.; Hauru, L.; Hummel, M.; Sixta, H., High- N-methylmorpholine-N-oxide. Macromol. Mater. Strength Composite Fibers from Cellulose-Lig- Eng. 2008, 293, 907-913. nin Blends Regenerated from Ionic Liquid Solu- [23] Wendler, F.; Todi, L.-N.; Meister, F., Thermosta- tion. ChemSusChem 2015, 8 (23), 4030-4039. bility of imidazolium ionic liquids as direct sol- [35] Asaadi, S.; Hummel, M.; Sixta, H. In Regenerated vents for cellulose. Thermochim. Acta 2012, cellulosic fiber from ionic liquid-waste cotton 528, 76-84. solution by dry-jet wet spinning, ACS spring [24] Zweckmair, T.; Hettegger, H.; Abushammala, H.; meeting, Denver, American Chemical Society: Bacher, M.; Potthast, A.; Laborie, M.-P.; Rosenau, Denver, 2015; pp CELL-138. T., On the mechanism of the unwanted acetylation [36] Michud, A.; Hummel, M.; Sixta, H., Influence of of polysaccharides by 1,3-dialkylimidazolium ace- molar mass distribution on the final properties of tate ionic liquids: part 1—analysis, acetylating fibers regenerated from cellulose dissolved in ionic agent, influence of water, and mechanistic consi- liquid by dry-jet wet spinning. Polymer 2015, derations. Cellulose 2015, 22 (6), 3583-3596. 75, 1-9.

75 LENZINGER BERICHTE 94 (2018) 67 – 76

[37] Hauru, L. J.; Hummel, M.; Michud, A.; Sixta, H., [43] Lenz, J.; Schurz, J., Fibrillar structure and defor- Dry jet-wet spinning of strong cellulose fila- mation behavior of regenerated cellulose fibers I. ments from ionic liquid solution. Cellulose 2014, Elementary fibrils and deformation. Cellul. Chem. 21 (6), 4471-4481. Technol. 1990, 24, 679-692. [38] Froschauer, C.; Hummel, M.; Iakovlev, M.; [44] Lenz, J.; Schurz, J.; Wrentschur, E., On the elonga- Roselli, A.; Schottenberger, H.; Sixta, H., Separa- tion mechanism of regenerated cellulose fibers. tion of Hemicellulose and Cellulose from Wood Holzforschung 1994, 48, 72-6. Pulp by Means of Ionic Liquid/Cosolvent Sys- [45] Schild, G.; Liftinger, E., Xylan enriched viscose tems. Biomacromolecules 2013, 14, 1741-1750. fibers. Cellulose 2014, 21 (4), 3031-3039. [39] Roselli, A.; Hummel, M.; Monshizadeh, A.; [46] Zhang, H.; Tong, M., Influence of hemicelluloses Maloney, T.; Sixta, H., Ionic liquid extraction on the structure and properties of Lyocell fibers. method for upgrading eucalyptus kraft pulp to Polym. Eng. Sci. 2007, 47 (5), 702-706. high purity dissolving pulp. Cellulose 2014, 21 [47] Treiber, E., Einfluss der Hemicellulosen auf die (5), 3655-3666. Weiterverarbeitung von Chemiezellstoffen. Papier [40] Northolt, M. G.; den Decker, P.; Picken, S. J.; 1983, 37 (12), 591-600. Baltussen, J. J. M.; Schlatmann, R., The Tensile [48] Kuzmina, O.; Bhardwaj, J.; Vincent, S. R.; Strength of Polymer Fibres. In Polymeric and In- Wanasekara, N. D.; Kalossaka, L. M.; Griffith, organic Fibers, Springer Berlin Heidelberg: Ber- J.; Potthast, A.; Rahatekar, S.; Eichhorn, S. J.; lin, Heidelberg, 2005; pp 1-108. Welton, T., Superbase ionic liquids for effective [41] Röder, T.; Moosbauer, J.; Kliba, G.; Schlader, S.; cellulose processing from dissolution to carbonisa- Zuckerstätter, G.; Sixta, H., Comparative char- tion. Green Chem. 2017, 19 (24), 5949-5957. acterisation of man-made regenerated cellulose fibres. Lenzinger Ber. 2009, 87, 98-105. [42] Lenz, J.; Schurz, J., Fibrillar structure and deformation behavior of regenerated cellulose fibers I. Methods of investigation and crystallite dimensions. Cellul. Chem. Technol. 1990, 24, 3-21.

76 LENZINGER BERICHTE 94 (2018) 77 – 84

Man-Made Cellulosic Fibers via the Viscose Process – New Opportunities by Cellulose Carbamate

R. Protz, G. Weidel, and A. Lehmann*

Fraunhofer Institute for Applied Polymer Research, Geiselbergstr. 69, 14476 Potsdam, Germany * Contact: [email protected]

Abstract

Viscose fibers have an ever increasing market share with over 5 million tons p.a.. Main application area for viscose fibers is the textile industry. Nevertheless the broad variability of the viscose process allows realizing even technical man-made cellulose fibers, like Supe 3 tyre cord yarn with tenacities of more than 50 cN/tex. This broad variability of fiber qualities is not reached by any other cellulose shaping process. However since decades there is a strong focus on decreasing the input of highly toxic CS2 for the viscose fiber manufacturing and keeping the quality of spinning solutions at the same time, which allows the known processing behavior and resulting in a high fiber quality. The cellulose carbamate process is by far the most intense investigated option to substitute the viscose technology. By different reasons this didn´t happen up to today, despite the fact that the cellulose carbamate spinning was already shown on industrial scale. This paper discusses the opportunity of combining both spinning routes, carbamate and viscose. This enables to realize fiber properties known from the viscose process and at the same time a significant lower 2CS input for so produced man-made cellulose fibers. The paper will give an overview about the technology, different fiber properties and structural characterization of the same by varying the processing conditions.

Keywords: man-made cellulose filaments, cellulose carbamate, viscose, cellulose

Introduction

Viscose fibers have an ever increasing market share This fact as well as the discussed cotton-gap in litera- with over 5 million tons p.a. [1]. Main application area ture [3] is the main driver to look into alternatives for for viscose fibers is the textile industry. This more than making man-made cellulosic fibers. The processing of 100 year old process discovered by Cross et al. [2] is cellulose from N-Methylmorpholine-N-oxide-Mono- still the dominating process to transfer the unspinable hydrate solution via an air-gap spinning method is the wood-based biopolymer cellulose into man-made fibers only further industrialized process to manufacture via the intermediate sodium cellulose xanthate. As man-made fibers in >200 kt-scale [4], giving man-made this transformation is a very complex process driven cellulose fibers called Lyocell (CLY). Owed to the dif- by physics and chemistry a broad variety of properties ferent dissolution and shaping conditions combined from resulting man-made cellulosic fibers is accessible with a non-derivatizing route, the achieved man-made by varying process parameters like spinning dope com- cellulosic fibers differ in their supermolecular struc- position (cellulose and alkali content, maturity, etc.) ture compared to viscose fiber (CV) leading to different and spinning conditions (coagulation conditions, draw- textile-physical properties [5], whereas the property ing factors, etc.). Despite the fact that the recycling window of CLY-fibers is not as versatile as it is known for the resulting sulphur containing by-products in the from CV-fibers. The viscose process allows even to viscose process became more and more efficient, it is produce technical filament yarns (Super 3), which are still a fact that much more than 1 million tons per year mainly used for rubber good reinforcement in particu- of the highly toxic carbon disulfide are needed to give lar fast running and run-flat tires. an output of 5 million tons of viscose fibers a year.

77 LENZINGER BERICHTE 94 (2018) 77 – 84

At least in pilot scale different alternative routes to Table 1: Concentration and amounts of cellulose carbamate realize man-made cellulose fibers were investigate and and sodium cellulose xanthate solutions to adjust the certain realized. Most of them following a non-derivatizing ratio in the spinning dope. route, whereas the dissolving pulp (in some cases addi- CC:Na-CX Cellulose Sodium cellulose tionally pre-treated by acid, alkali or enzymes) is dis- carbamate dope xanthate dope solved directly in a solvent and the resulting spinning dope is shaped via wet-spinning (e.g. Celsol [6]) or air- 33:67 1 part 9.5 wt.-% 0.7 parts 8.8 wt.-% 67:33 1 part 9.5 wt.-% 2.4 parts 10.1 wt.-% gap spinning technique (e.g. Ioncell [7]) to create man- made cellulose fibers. An example for a derivatizing route is the so called CarbaCell® process [8]. Cellulose Cellulose carbamate based solution is modified with urea to result in an alkali soluble The cellulose carbamate was mixed with water 1:1 by cellulose carbamate (CC), which can be shaped very weight and stored over night at 6 °C. The dissolution similar to conditions applied for creating man-made lye was introduced in a dissolving vessel equipped cellulose fibers via the viscose route. Despite the fact with rotor/stator geometry and cooled down to -4 °C. that cellulose carbamate is a derivative like cellulose After reaching the temperature the cellulose carbamate xanthate, the main difference is indeed that cellulose was given to the dissolution lye under stirring and dis- xanthate is a polyelectrolyte, whereas cellulose carba- solved for 120 min at 0 °C. The so obtained cellulose mate is a non-ionic cellulose derivative. So the influ- carbamate solution had a cellulose content of 9.5 wt.-% ence of precipitation conditions, especially electrolyte and an alkali content of 7 wt.-%. content like salt concentration as well as kind of salt is much less pronounced in the CarbaCell® process com- Sodium cellulose xanthate based solution pared to the viscose process. As both derivatives are The pulp was cut into sheets (18 x 22 cm) and placed soluble in alkali, we investigated in the present paper into the steeping chamber of a Blaschke plant. Alkali the possibility to combine both cellulose derivatives in lye of 18 wt.-% in concentration was slowly given solution, shaping the same into man-made cellulosic from bottom to top into the alkalization chamber. The fibers and characterizing the resulting filaments. steeping was performed for 50 minutes at 35 °C. The sheets of alkali cellulose (AC) were than pressed to Materials and Methods obtain the press out factor of 2.9 The resulting alkali cellulose consists of 36.9% cellulose and 14.5% Raw materials alkali and was disrupted by the defibrator of the For performing the investigations a cellulose carbamate Blaschke plant. The so prepared alkali cellulose was with a DPCuoxam of 254 was used having a nitrogen con- aged at 35 °C to reach a DPCuoxam of 340. tent of 2.1%. This cellulose carbamate was synthesized Xanthation took place in a so called baratte. Therefore by the method described in [9]. For the viscose dope the alkali cellulose was introduced in the baratte and preparation a softwood pre-hydrolysis kraft pulp, having vacuum was applied. Carbon disulfide (28 wt.-% based a DPCuoxam of 611 and an -cellulose content of 92.3%, on the cellulose) was flushed into the de-pressurized was used. For the xanthation reaction carbon disulfide reaction chamber, which lead to loss of vacuum. The with a quality of 99.9% from Fisher was used. xanthation took place for 90 min at 28 °C. In that time the pressure decreases again to the starting value, in- Methods dicating the complete usage of carbon disulfide for the xanthation reaction. Spinning solution preparation The resulting, yellowish sodium cellulose xanthate had A cellulose content of 8.5 wt.-% and an alkali content a -value of 51.4 and was dissolved by pre-cooling of 7 wt.-% was addressed for all prepared spinning so- the dissolution lye to 6 °C and introducing the sodium lutions by varying the ratio between cellulose carbamate cellulose xanthate to the dissolving equipment (rotor/ and sodium cellulose xanthate for the spinning dope as stator dissolving geometry of 5l volume). After final well as for shaping cellulose carbamate and sodium cel- addition of the sodium cellulose xanthate the further lulose xanthate themselves. The mixed cellulose solu- dissolution took place for 120 min at 6-8 °C. tions were prepared by combining cellulose carbamate After combining the sodium cellulose xanthate and solution with a cellulose content of 9.5 wt.-% and ad- cellulose carbamate solution in appropriate amounts justing the cellulose content in the viscose dope to (Table 1) to realize the addressed ratio, the spinning 8.8 wt.-%, respectively 10.1 wt.-% in appropriate amounts dope was filtered by a 20 µm metal fleece filter. to realize the later discussed ratios (Table 1).

78 LENZINGER BERICHTE 94 (2018) 77 – 84

The maturity was realized by storing the spinning using demineralized water at 60 °C. The washed yarn dope at 25 °C for 20 hours. was dried under isometric conditions using drying rollers at 80 °C and finally wound on bobbins. Spinning of filament yarn Jet-stretch and draw factor were varied by adjusting For wet-spinning of the resulting spinning solutions the throughput as well as spinning speed to realize the the spinning dope was tempered at 20 °C during spin- spinning parameters given in Table 2 . ning and transported by a spinning pump (0.6 cm³ per rotation) to the spinning nozzle. As spinning nozzle a Table 2: Applied jet-stretch and draw factors during the spinneret with 128 holes, each hole having a diameter of spinning process. 70 µm was used. The spinning bath consists of 80 g/l sulfuric acid and 120 g/l sodium sulfate in water at Jet-stretch Draw-factor 40 °C. The resulting bundle of never-dried filaments 0.6 1.1 0.7 was drawn from the nozzle by a take-up roller, which 0.9 delivered the yarn to the decomposition bath, which 0.7 1.3 contained 20 g/l sulfuric acid in water at >90 °C, 1.5 followed by washing the yarn on nelson-type rollers 1.7

Figure 1: Schematic view of the spinning line for filament yarns at Fraunhofer IAP.

Analytical methods (λ = 0.15418 nm) monochromized by a Ge(111)-primary monochromator and operated at 40 kV and 30 mA. Rheology Distribution of the (004) lattice plane (chain direction Oscillation rheology of all solutions was measured on peak) at a diffraction angle 2θ at ~34-35° was measured an Anton Paar MCR 102 rheometer with a plate and by a φ-spinner-scan over φ-angle range from 60-120° plate geometry. A plate with 50 mm diameter was used with steps of 0.2° and a scanning speed of 0.067°/min and the gap was set to 1 mm. The viscoelastic domain (assumed maximum of distribution to be 90°). By fol- was determined by performing a dynamic amplitude lowing the orientational parameters of the crystallite, sweep and a stain between 0.1% and 100%. A strain of the main axis orientation was calculated in order to 1%, which is within the linear viscoelastic regime, was describe the peak reflex of the (004) plane: chosen for time sweep measurements. In order to keep (180°– FWHM) the moisture content constant the plate edge was sealed OG(004) = (180°) with paraffin oil. The time sweep (frequency 1 Hz and strain amplitude 0.01) was run at 25 °C for 10 hours. SEM The SEM was used to study the morphology of cross WAXS – Quantitative analysis of the crystalline orien- sections. The SEM-micrographs were taken by a tation of cellulose fibers JSEM 6330F (Jeol, Japan) at an acceleration voltage The fibers were parallel- fixed onto a special sample of 5 kV. Cross sections were prepared by fracturing holder in order to analyze the crystallite orientation. fibers in liquid nitrogen and finally, a thin Platinum X-ray measurements were carried out by a two-circle layer (thickness 5 nm) was deposited by sputtering on D5000 X-ray diffractometer (Bruker-AXS) in sym- top of the surface and cross section to avoid electrical metrical transmission geometry using CuKα-radiation charging by the electron beam.

79 LENZINGER BERICHTE 94 (2018) 77 – 84

Fiber testing All produced filaments were stored at 65% relative humidity and 20 °C for at least 24 h for sample con- ditioning.

Titer determination The titer of the single filament was determined by a vibroskop (Lenzing) as average value of 20 samples.

Mechanical parameters of filaments Mechanical properties (tenacity, elongation at break and modulus [sloop in the range of 0.05-0.25% elonga- Figure 2: Oscillatory time sweeps at 25 °C for a carbamate tion]) of samples were measured using a UNIVERSAL solution and a viscose/carbamate-mixture (33/67, wt./wt.) TEST MACHINE ZWICK Z 020. The clamp distance with strain amplitude 0.01 and frequency 1 Hz. was 20 mm and the testing velocity 10 mm/min by a pre-tension of 0.1 cN. The measurement of 20 samples was averaged. Results and Discussions

Rheology of spinning dopes The stability of carbamate and viscose solutions at 25 °C processing temperatures is an important parameter for processability of the corresponding spinning dopes. A substantial difference between cellulose carbamate (CC) and sodium cellulose xanthate is the chemical stability of the CC for months in the dry state [10]. Figure 3: Oscillatory time sweeps at 15 °C for a carbamate And compared to a sodium cellulose xanthate solution, solution and viscose/carbamate-mixtures (33/67 and 67/33, where the maturity is a very important factor for the wt./wt.) with strain amplitude 0.01 and frequency 1 Hz. with resulting fiber properties, a carbamate solution does not strain amplitude 0.01 and frequency 1 Hz. require any ripening time [11]. Due to the effects of hydrolysis of the carbamate groups carbamate solutions Figure 3 demonstrates the changes in rheology and tend to gelify earlier and therefore should be kept cold appearance of mixtures from sodium cellulose xanthate [12].Ongoing changes in the physical and chemical and cellulose carbamate containing different amounts structure in both cases lead to a network build-up up to of cellulose carbamate being stored at 15 °C. It can be the point of gelification. Pinnow et al. [13] used this net- noticed that initially all solutions are dominantly viscous work formation and showed the thermal induced preci- (G’

80 LENZINGER BERICHTE 94 (2018) 77 – 84

investigated spinning solution. The set of parameters A much more pronounced effect on the textile-physical is given in Table 2 (page 79). properties for shaping at a ratio of a spinning dope, which consists of a mixture CC:Na-CX of 33:67, can be detected for the draw factor (Figure 5) by keeping the jet-stretch constant. The alignment of the cellulosic polymer chains along the fiber axis is increased by realizing higher draw factors. This results in a higher orientation of the crystalline as well as non-crystalline area of the filaments giving higher values for tenacity and modulus as well as lower values for elongation at break. Despite the fact, that a drawing by factor 1.7 is possible under stable spinning conditions, the further increase in textile-physical properties is limited in comparison to a draw factor of 1.5. Figure 4: Influence of jet-stretch at a constant draw factor of Figure 6 shows an overview of the textile-physical 1.1 on textile-physical properties for shaping cellulose car- properties of filaments, which were spun by varying bamate and sodium cellulose xanthate at a ratio of 67:33. the ratio of cellulose carbamate to sodium cellulose xanthate in the spinning dope and shape the same at At a constant draw-factor, the jet-stretch has a minor given spinning parameters. Comparing the shown values influence on the textile-physical properties as shown for filaments for using cellulose carbamate as derivative in Figure 4. The tenacity is nearly not influenced. for the shaping process lower values for tenacity and When the jet stretch becomes higher than 0.6 it can be elongation at break in comparison for the usage of seen that the value for the modulus increases, whereas sodium cellulose xanthate as derivative by using men- the value for the elongation at break decreases. tioned spinning conditions can be detected, whereas minor differences in modulus are observed. While the values for elongation at break are increasing by increasing the part of sodium cellulose xanthate in the spinning solution the values for tenacity of resulting filaments are already on the same level like those prepared from a pure viscose based dope by using 1/3 sodium cellulose xanthate and 2/3 cellulose carbamate in the spinning solution. Interestingly there is a slight trend of realizing the highest values for the modulus by using cellulose carbamate and sodium cellulose xanthate at the same time in the spinning dope, independent from the ratio. Figure 5: Influence of draw factor at a constant jet-stretch of By the achieved results it can be clearly seen, that an 0.7 on textile-physical properties for shaping cellulose carba- partly usage of sodium cellulose xanthate in a cellulose mate and sodium cellulose xanthate at a ratio of 33:67. carbamate based spinning solution enables a more balanced ratio between the values for tenacity and elongation at break and so mimicking more the textile- physical properties of a viscose based man-made fiber (Figure 7, page 82). To realize this, same spinning bath composition was used and jet-stretch and draw factors had to be adjusted.

Figure 6: Influence of the ratio of cellulose carbamate (CC) to sodium cellulose xanthate (Na-CX) in the spinning dope for shaping the same at a jet-stretch of 0.6 and a draw factor of 1.1.

81 LENZINGER BERICHTE 94 (2018) 77 – 84

composition described in Figure 7 the carbon disulfide input could be decreased down to 9.3 wt.-% (meaning 93 kg CS2) per ton man-made cellulose fiber and still have the shaping options of a cellulosic polyelectrolyte structure.

Investigation of fiber structure On filaments produced by spinning under variation of the draw factor and differing in their ratio of used cellulose carbamate to sodium cellulose xanthate some structural parameters by X-ray methods and electron microscopy were estimated. Figure 7: Comparison of textile-physical properties of filaments produced by the viscose and carbamate process as well Orientational parameters of crystalline phase as the combined process applying the same spinning bath As shown in Figure 8 the distribution of the (004) lattice conditions. plane (chain direction peak) at a diffraction angle 2θ at ~34-35° by a φ-spinner-scan over φ-angle range Considering an input of 28 wt.-% of carbon disulfide from 60-120° for a draw factor of 1.1 (right) and those based on cellulose for realizing 1 ton of man-made for a draw factor of 1.5 (left). In Table 3 the values of the cellulose fiber (meaning 280 kg CS2), by using the corresponding degree of orientation (OG(004)) are listed.

Azimuthal angle  [deg] Azimuthal angle  [deg] Figure 8: Distribution of the (004) lattice plane of produced filaments as a function of spinning dope composition and draw factor.

For applying a low drawing almost no differences for OG(004) values was determined. The strongest increase the values of the OG(004) can be detected, which cor- can be seen when cellulose carbamate and sodium responds to the measured E-Moduli (Table 3). By in- cellulose xanthate are combined in the spinning dope creasing the drawing factor to 1.5 an increase of all and shaped together.

Table 3: Degree of orientation of the (004) lattice plane (OG(004)) and the corresponding values for the E-Moduli of produced filaments as a function of spinning dope composition and draw factor.

JS: 0.7 DF: 1.1 JS: 0.7 DF: 1.5

Filament sample OG(004) E-Modulus [cN/tex] OG(004) E-Modulus [cN/tex] CC:CV 67:33 0.873 970 ±83 0.901 1472 ±96 CC:CV 33:67 0.875 895 ±47 0.897 1517 ±72 CC:CV 100:0 0.872 934 ±38 0.888 1329 ±85

82 LENZINGER BERICHTE 94 (2018) 77 – 84

Morphological properties

Figure 9: Electron micrographs of cross-section from different cellulosic man-made fibers; a) CC:CV 0:100, b) CC:CV 33:67, c) CC:CV 67:33, d) CC:CV 100:0.

In Figure 9 the picture (a) and (d) depict a cellulose In contrast to that such lobulated morphology cannot be man-made filament according to the viscose process detected for a cellulose man-made filament realized (a) and the cellulose carbamate process (d). From pic- via the carbamate route. As cellulose carbamate is not ture (a) a slightly lobulated morphology can be seen, a polyelectrolyte the structure formation during the which is dedicated to the structure formation of the poly- precipitation step in the spinning bath is less complex electrolyte sodium cellulose xanthate during the preci- and more dedicated to the neutralization of the solvent pitation step in the sodium sulfate containing acidic (alkali lye) by the acidic spinning bath. spinning bath. The diffusion of the acid from the spin- Comparing pictures (b) and (c) it can be seen, that as ning bath in the spinning rays of the viscose dope lead long as sodium cellulose xanthate is the majority of to neutralization and finally acidification of the same. the fiber forming cellulosic polymer in spinning solu- During this process sodium sulfate is formed as a by- tion, the slightly lobulated morphology can be found product of the reaction of the alkali in the spinning in the corresponding filament, too. This lobulated dope and the acid in the spinning bath resulting in a morphology cannot be seen anymore as soon as cellu- dehydration and densification of the sodium cellulose lose carbamate becomes the majority of the cellulosic xanthate. Thereby gradients in the concentration of polymers in the filament (c). the acid as well as of the salt content in- and outside of the spinning rays causing a centripal shrinkage of the just formed gel by osmotic and synaeresic processes [14], which leads to non-uniform coagulation and so in a lobulated cross-section shape.

83 LENZINGER BERICHTE 94 (2018) 77 – 84

Conclusions References

From the found and discussed results it can be con- [1] Global and China Viscose Fiber Industry Report, cluded, that shaping cellulose carbamate and sodium 2013-2016 cellulose xanthate at the same time from alkaline solu- [2] C. F. Cross E. T. Bevan C. Beadle; Berichte der tion is possible. The ratio of both cellulose derivatives deutschen chemischen Gesellschaft, 26, 1893 can be varied in a broad range. Using the discussed [3] F. M. Hämmerle; Lenzinger Berichte, 89, 2011, 12 precipitation conditions, filaments with very similar [4] Lenzing AG, Annual Financial Report 2014, 2015 textile-physical properties compared to viscose filaments [5] T. Röder, J. Moosbauer, K. Wöss, S. Schlader, can be realized. It was found that, in contrast to a G. Kraft; Lenzinger Berichte, 91, 2013, 7 cellulose carbamate solution, the solution stability [6] D. Ciechanka, D. Wawro, W. Steplewski, J. Kazi- remarkably increases when partly sodium cellulose mierczak, H. Struszczyk; Fibres and Textiles in xanthate is added. Eastern Europe, 13, 2005, 19 Anyhow, there are not-answered questions like the in- [7] L. K. J. Hauru, M. Hummel, A. Michud, H. Sixta; fluence of nitrogen containing by-products in the Cellulose, 21, 2014, 4471 recycling stream of the sulphur containing by-products [8] M. Voges, M. Brück, H.-P. Fink, J. Gensrich; in a viscose plant, the origin of cellulose carbamate as Cellulosic Man-made Fibres in the new Millenium, well as a deeper investigation on fiber structure and Stenungsund, 2000 textile-physical properties and further improvement [9] US 5,906,926 of the same. [10] H.-P. Fink, J. Ganster, A. Lehmann; Cellulose, 21, 2014, 31 Acknowledgements [11] K. Götze; Chemiefasern nach dem Viskosever- fahren, Springer-Verlag, Berlin-Heidelberg-New The authors would like to thank Dr. Andreas Bohn, York, 1967 from Fraunhofer IAP, for performing X-Ray measure- [12] K. Ekman et al.; Modification and Hydrolysis in ment and electron microscopy. Cellulose Structure, John Wiley & Sons, New York, 1986, 131 [13] M. Pinnow et al.; Macromolecular Symposia, 262, 2008, 129 [14] A. Herzog; Textile Forschung, 8, 1926, 87

84 LENZINGER BERICHTE 94 (2018) 85 – 94

Influence of Carbonization Aids on the Manufacture of Carbon Fibers based on Cellulosic Precursors

Johanna M. Spörl1, Frank Hermanutz1, and Michael R. Buchmeiser1,2,*

1 Deutsche Institute für Textil- und Faserforschung (DITF) Denkendorf, Körschtalstr. 26, D-73770 Denkendorf, Germany 2 Lehrstuhl für Makromolekulare Stoffe und Faserchemie, Institut für Polymerchemie, Universität Stuttgart, Pfaffenwaldring 55, D-70550 Stuttgart, Germany * Corresponding author: Tel. +49(0)711-685 64075, [email protected]

Abstract

The application of a new sulfur-based carbonization aid in the manufacture of cellulose-based carbon fibers (CFs) was studied. A process for stabilization was established providing undamaged fibers suitable for continuous carbonization. High mechanical strength cellulose-based CFs were manufactured at comparably low temperatures and in high yield (37% = 83% of theory). Maximum tensile strengths and Young’s moduli up to 2 and 84 GPa were obtained at 1400 °C. A well-known phosphorous-based carbonization aid was studied comparatively. Both, the precursor and the CFs were characterized via elemental analysis, wide-angle X-ray scattering (WAXS), scanning electron microscopy (SEM), and tensile testing. Additionally, thermogravimetric analysis coupled with mass spectrometry/infrared spectroscopy (TGA-MS/IR) was used to disclose differences in pyrolysis and structure formation.

Keywords: cellulose; carbon fibers; TGA-MS; TGA-IR; WAXS

Introduction

Cellulose is the oldest known carbon fiber (CF) pre- to degradation reactions and the formation of carbon cursor material. After major research activity on cellu- containing volatiles (like CO2, CO, alcohols, ketones losic precursors in the 1950-1970s, the research in this and other carbon-containing low-molecular-weight field was abandoned due to disadvantages such as high substances), the total mass loss might be up to 90%. [1] production costs at low yields and the much more prom- Different strategies for the stabilization of rayon fibers ising findings with polyacrylonitrile (PAN)-based pre- prior to their carbonization have been reported allowing cursors. Today, interest in cellulosic CF precursors has for carbonization yields of 25-30 wt.-%: the use of arisen, for both economic and ecological reasons. [1-8] low heating rates during pyrolysis up to 400 °C [9-11], Man-made regenerated cellulose fibers, like viscose oxidative pre-treatment [12-16], the use of a reactive or Lyocell possess well-defined dimensions and high atmosphere [17] or the application of carbonization purity. Therefore, they could be used as precursors for aids, including catalysts for dehydration (AlCl3, bo- high-performance CFs when processed as continuous rax, boronic acid, KHCO3, NaH2PO4, ZnCl2, NH4Cl, multi-filaments. The mechanical performance of CFs ammonium phosphates and sulfates) and crosslinking also strongly depends on controlled processes during reagents, e.g., urea [5,14,18-27]. We recently reported pyrolysis and carbonization. The maximum theoretical on an ionic liquid-(IL)derived cellulose phosphonate carbon yield in carbonization of cellulose is 44.4 wt.-%, precursor fiber for CFs, which allows for high carboni- corresponding to the formal loss of five molecules of zation yields of 36% at 1400 °C [28]. An overview of water per anhydroglucose unit (AGU). However, due cellulose-based CFs is given in ref.[1] However the

85 LENZINGER BERICHTE 94 (2018) 85 – 94

influence of structure formation on mechanical prop- tion trials, a Gero vertical tube furnace was used. Fibers erties of cellulose based CFs was still not fully en- were continuously stabilized under nitrogen at 210 °C lightened. We recently presented a study on Rayon tire (RC/ATS) or 250 °C (RC/ADHP) for 50 minutes. cord- (RC-) derived CFs aiming on both, high carboni- Stabilized fibers were then carbonized at 1400 °C using zation yields and good mechanical properties of the different dwell times and fiber tensions. In course of final CFs. [29] Therefore, ammonium tosylate (ATS) these experiments, the preload force on the fiber was was tested as new sulfur-based carbonization aid in com- increased until the first filaments of the fiber thread parison to the well-known carbonization aid ammonium broke. For the trails without preload force, the fiber dihydrogenphosphate (ADHP). Carbonization trials tension was 0.5 cN/tex. Selected fibers were also car- were performed continuously to study the influence of bonized in a Gero horizontal tube furnace consisting of carbonization temperature, dwell time and fiber tension two zones, a low temperature zone (carbonization at during carbonization and the results were correlated up to 750 °C, dwell time at Tmax = 4.5-5 min) and a with data obtained from discontinuous processing. In high temperature zone (carbonization at 1400 °C, dwell the present work, a closer look is taken on emission time at Tmax = 3.6 min). gases evolved during pyrolysis and carbonization of untreated and treated cellulose fibers. Furthermore, the Characterization influence of the two different carbonization aids on the Quantitative phosphorous and sulfur analysis was per- properties of the resulting CFs is discussed. formed by inductively-coupled plasma-optical emission spectroscopy (ICP-OES) on a Spectro Acros ICP-OES after acidic digestion. Calibration was accomplished Materials and Methods with aqueous ICP standards in the concentration range of 1 to 7.5 mg/L at λ = 214.900-214.940 nm (base Materials line: λ = 214.784-214.813 nm, 214.985-215.022 nm) A high-strength Rayon fiber RC (Cordenka 700, 1000 for phosphorous and λ = 180.714-180.762 nm (base filaments, 1840 dtex) was used. According to single line: λ = 180.620 -180.650 nm, 180.829-180.860 nm) filament fineness and tensile testing (own data), the for sulfur. fiber properties were: 18% elongation at break ε, Scanning electron micrographs (SEM) were recorded 53 cN/tex tenacity σ, 1290 cN/tex Young’s modulus E on a Zeiss Auriga field emission scanning electron mi- and a fineness T of 2.1 dtex. (NH4)H2PO4 was pur- croscope. Samples were sputtered with Pt/Pd prior to chased from ABCR GmbH; ammonium tosylate was analysis. synthesized by adding 1.3 equiv. of a 25 wt.-% aqueous Mechanical fiber properties were measured according solution of ammonia to an aqueous solution of p-toluene to EN ISO 5079 by a Textechno Favimat tensile tester. sulfonic acid, steering the mixture for 1 h and sub- Fineness was obtained as weight in g per 10,000 m in sequently removing the solvent in vacuo. For compara- length (dtex). tive analysis, a commercial PAN-based CF T300 For WAXS measurements, a Rigaku D/Max Rapid II (Toray Industries Inc.) was used. The tensile proper- was used at 40 kV and 30 mA with Cu Kα radiation ties (single filament test) were ε = 1.3%, σ = 2.5 GPa (λ = 1.54 Å). A shine monochromator and an image and E = 190 GPa. plate detector were used. The scanning rate was 0.2°/min; the scanning step was 0.1°. All fibers were Application of the carbonization aids aligned in a fiber sample holder. Pearson-VII curves The application of the carbonization aids was accomp- were fitted to the crystalline signals and the crystallite lished continuously by dipping the fiber in a 0.4 M size (t) was calculated according to the Scherrer’s solution of the respective carbonization agent (retention equation [30] time: approx. 1.5 s) and subsequently drying the fibers t = Kλ on godets (80 °C) and a heating channel (100 °C). (B·cos θb), Finally, the fibers were wound on a spool. where B is the width at half-height of the reflection, K Carbonization is the Scherrer factor, which depends on the crystallite For discontinuous carbonization trials, fibers were car- shape, and θb is the Bragg-angle. For CFs, the degree bonized under nitrogen in a Gero HTK 8GR/22-1G of preferred relative orientation of the crystallites P.O. high-temperature batch type furnace applying a heat- was calculated according to: ing rate of 1 K/min below 350 °C and 5 K/min above this temperature. The final carbonization temperature P.O. = 180°- B varied from 400 to 2200 °C. For continuous carboniza- 180°

86 LENZINGER BERICHTE 94 (2018) 85 – 94

Thermogravimetric analysis coupled to mass spectro- Table 1: TGA results of untreated fibers and fibers containing metry and IR-spectrometry (TGA-MS/IR) was accom- different amounts of ADHP and ATS characterized by their plished in a helium atmosphere at 10 K/min on a phosphorous or sulfur content (rt - 1400 °C, 10 K/min, helium). NETSCH STA 449 F3 coupled to a QMS 403 C Aelos and a Bruker Tensor 27 FTIR. For all comparisons, entry phos- sulfur onset T residual mass [wt.-%] of decompo- at 1400 °C identical sample masses were used. phorous [wt.-%] sition [°C] [%] Results and Discussion untreated 0 – 325 17 RC/ADHP1 0.6 – 239 33 Dehydration and stabilization of cellulose RC/ADHP2 1.0 – 224 38 with carbonization aids RC/ADHP3 1.8 – 218 34 The fibers were finished with different amounts of ADHP or ATS from aqueous solution, dried and sub- RC/ATS1 0.4 210 32 jected to TGA in comparison to the untreated fiber RC/ATS2 1.0 210 37 (Figure 1, Table 1). All thermograms showed a first RC/ATS3 2.3 204 35 mass loss around 100 °C which could be attributed to the evaporation of physically absorbed water and a OH-groups of cellulose, resulting in a good leaving second main mass loss step. The onset temperature of group in order to dehydrate the AGU. Thereby cellu- this main mass loss step was at 320 °C for the untreated lose is stabilized. In order to proof this presumption, the RC, while it reduced to 224 °C for the fiber treated emission gases evolved during heat treatment of the with ADHP (1 wt.-% P) and 210 °C for the fiber treated fibers were studied by IR spectroscopy in order to gain with ATS (1 wt.-% S). The residual mass in TGA at information on the mechanism of pyrolysis. The emission 1400 °C increased from 17% for the untreated fiber products H2O, CO2, CO and CH4 as well as the evolu- sample to 38 for the fibers finished with ADHP tion of gases with functional groups like C=O, C-O and (1 wt.-% P) and 37% for ATS treated fibers (1 wt.-% S), CH2 could be detected, which belong to decompositi- respectively. This corresponds to 86 and 83% of the on products like aldehydes, ketones, carboxylic ester theoretical carbon yield and an increase of the carboni- and alcohols, e. g. (hydroxy)acetaldehyde, acetone, zation yield of more than 100% compared to untreated furane derivatives and levoglucosan. The emission ga- fibers. An optimum concentration of ADHP and ATS ses considered in this study are summarized in Table 2. with respect to the residual mass in TGA was found at 1.0 wt.-% of P or S, respectively (Table 1). Table 2: Evaluation and assignment of emission gases by IR.

decomposition mode band range1 baseline2 product / [cm-1] [cm-1] functional group

H2O νs / νas 3858-3850 3858-3850

CO2 νas 2396-2265 2396-2265

CO νs 2144-2000 2144-2000

CH4 νs / νas 3024-3003 3024-3003

C=O νs 1840-1626 1840-1626

C-O νs 1209-1143 1209-1143

CH2 νs / νas 3002-2859 3002-2859 Figure 1: TGA curves of fibers treated with ADHP (1wt.-% P) 1 band range in the spectrum used for integration, or ATS (1 wt.-% S) and untreated fibers (rt - 1400 °C, 10 K/min, 2 baseline for integration. helium).

It is known for ADHP that it decomposes to NH3 and H3PO4 during heat treatment. Thereby, it can act as a catalyst for dehydration reactions, thus making cellulose less prone to degradation reactions. [31] A similar mechanism is expected for cellulose and ATS, which forms the strong acid p-toluenesulfonic acid (pKa < 1) during thermal decomposition. This acid can esterify

87 LENZINGER BERICHTE 94 (2018) 85 – 94

C-O functionality were formed. With the ATS additive, a release of gas occurred over a longer temperature regime. Compounds with C=O and CH2 functionalities were formed to a significantly lower extent. The treated samples released more CO, especially in the range 400-800 °C. In contrast to the other emission products, the formation of CH4 occurred in the range 400-800 °C. The lowest amount of CH4 was released from the sample treated with ADHP. By contrast, the sample treated with ATS gave rise to the highest emission of CH4, because CH4 could also be formed from the additive itself. In contrast to the formation of tarry products from b pure cellulose,[31] even at low temperatures stable fibers formed, which were less prone to decomposition reactions. The suppression of tar formation is not only responsible for the increase of carbonization yield but also has a very positive effect on the final technical implementation because it prevents the fusing of filaments and the contamination of the furnace during carbonization. For continuous processing, stabilization of the precursor fibers equipped with ADHP (1 wt.-% P) or ATS (1 wt.-% S) was studied by isothermal TGA under N2 at the respective onset temperature of decomposition. After 55 min, mass losses of 37 and 30% were found at the respective stabilization temperature (Figure 3). Note that the release of three equivalents of H2O per AGU would correspond to a mass loss of 33%. The formation of small but detectable amounts of CO and CO2 during stabilization was revealed in both cases Figure 2: a) Normalized gas phase IR spectra of cellulose by a simultaneous analysis of the evolved gases via fibers (at T = 339 °C) and cellulose fibers treated with ADHP MS (Figure 3). (T = 246 °C) and ATS (T = 220 °C), respectively; b) release rates of selected emission gases during TGA of untreated cellulose fibers and cellulose fibers treated with ADHP and ATS as determined via IR spectroscopy.

Figure 2a depicts the IR spectra taken at the respective main mass loss step of the untreated and treated fibers. For the untreated fiber, the H2O band was less pronounced whereas the C-H and C-O bands showed a higher intensity than for the treated fibers. Thus it can be seen, that the application of ADHP or ATS acted catalytically in the dehydration reaction [31-35] during pyrolysis resulting in an increased emission of H2O. Moreover, less CO2 was formed in the main mass loss step for the fiber treated with ATS. The release rates of the emission products are depicted in Figure 2b as a function of temperature. The temperatures of maximum release of the gases mostly followed the order of the onset temperatures of the TG signals. For the treated samples, the release of water occurred earlier and in larger amounts than for the untreated fiber Figure 3: Isothermal TGA-MS of RC/ADHP at 250 °C (top) sample. Moreover, less CO2 and less compounds with and RC/ATS at 210 °C (bottom).

88 LENZINGER BERICHTE 94 (2018) 85 – 94

Elemental analysis revealed similar C, H and N con- Table 3: Elemental analysis of continuously stabilized fibers tents of the stabilized fibers irrespective of the car- derived from RC/ATS and RC/ADHP. bonization aid used for impregnation (Table 3). Fi- precursor EA [wt.-%] bers consisted of almost 60 wt.-% C while about C H N S 2 mol of oxygen per AGU remained in the stabilized RC/ATS 59.8 4.5 0.5 1.2 fiber. RC/ADHP 58.1 4.4 0.8 —

Continuous carbonization of the precursor fibers carbonization was performed at at different fiber tensions 1400 °C while applying different amounts of tensile force Continuous carbonization of an RC fiber containing on to the fiber. A significant enhancement of the mecha- ATS (1 wt.-% S) and a RC fiber containing ADHP nical properties, especially Young’s modulus, could be (1 wt.-% P) were performed at 1400 °C. A continuous achieved by increasing the tensile force (Figure 4). carbonization of pure RC without any additive was However, for the RC/ATS precursor a higher tensile omitted, because in a discontinuous carbonization trial, force could be applied during carbonization experi- these fibers already led to very poor mechanical proper- ments, which ultimately resulted in superior mechani- ties and carbon yields. After continuous stabilization cal properties in the CF compared to RC/ADHP.

Figure 4: Mechanical properties of RC/ADHP- (left) and RC/ATS derived CFs (right) from continuous carbonization at 1400 °C in dependence of dwell time (6, 12, 24 or 48 min) and fiber tension.

With respect to the maximum tensile force which could during processing and thus reached better mechanical be applied before filament breakage the optimum dwell properties (Figure 4). Cellulose-based CFs with reasona- times for carbonization were 12 min for RC/ATS and ble mechanical properties (tenacity of 2.0 GPa, Young’s 24 min for RC/ADHP (Table 4, page 90). RC/ATS re- modulus of 84 GPa, elongation of 2.4%) were obtained sisted a 40% higher tension than RC/ADHP fibers with optimized conditions.

89 LENZINGER BERICHTE 94 (2018) 85 – 94

Table 4: Experimental conditions during continuous carbo- tances d002 were found for both precursor types, with nization at 1400 °C of the stabilized fibers. slightly lower values for RC/ATS-derived CFs at higher fiber tension. Comparing the two precursor types car- precursor dwell time maximum fiber bonized at dwell times of 12 and 24 min, the crystallite [min] tension1 [cN/tex] height Lc was lower for RC/ADHP. The crystallite length RC/ATS 6 1.6 La increased with the tensile strain applied during car- RC/ATS 12 2.6 bonization for both precursor types. Although RC/ADHP RC/ATS 24 1.8 could only withstand a lower tensile strain than RC/ATS, RC/ADHP 6 1.2 higher L values were reached. How-ever, this did not RC/ADHP 12 1.3 a affect the mechanical properties of RC/ADHP (see RC/ADHP 24 1.6 Figure 4). The applied force also resulted in an increase RC/ADHP 48 1.1 of the crystallites’ preferred orientation, P.O. In the 1 maximum fiber tension before filament breakage during continuous process without application of tensile force carbonization. P.O. was about 50%, while with tensile force, a maximum P.O. of 63% for RC/ATS- and 57% for RC/ADHP- Figure 5 depicts the crystalline parameters determined derived CFs was reached. This increase in P.O. correlated via WAXS analysis of RC/ADHP- and RC/ATS-derived with the increase in Young’s modulus. For comparison, CFs, which were processed at different dwell times, a commercial PAN-based CF T300 (Young’s Modulus: as a function of fiber tension during the continuous 190 GPa, probably carbonized at a temperature of carbonization at 1400 °C. Comparable interlayer dis- 1200-1400 °C), showed a P.O. value of 81%.

Figure 5: Interlayer distances, crystallite sizes and preferred orientation of RC/ADHP- (left) and RC/ATS-derived CFs (right) from continuous carbonization at 1400 °C and variable dwell times (6, 12, 24 or 48 min) and fiber tensions; T300: values obtained from a commercial PAN-based CF (Toray Industries Inc.).

The elemental composition of the CFs with respect In graphite-like structures elements other than carbon to the dwell time was also considered. Irrespective and hydrogen usually tend to accumulate between the of the dwell time, higher C contents of ca. 97 wt.-% graphite layers rather than to become incorporated were found with RC/ATS, in comparison to 92-93 into the layer. Therefore phosphorous and oxygen in wt.-% for RC/ADHP-derived CFs. Neither S nor RC/ADHP-based CFs seem to interfere with layer sta- H were found, but both CF types contained about cking, thus leading to smaller Lc values. The growth 1 wt.-% N. of La is instead favored already at low fiber tension.

90 LENZINGER BERICHTE 94 (2018) 85 – 94

Fiber morphology and fiber shape res with diameters of up to 75 nm were visible at high SEM images of RC-derived CFs from continuous car- image resolution, and yet, ultimate tensile strengths of bonization are depicted in Figure 6. After carbonizati- more than 1 GPa were reached. No pores were found on, the fibers retain the varying and irregular cross- in high resolution images at dwell times of 6 and 12 sections, which stem from the viscose manufacturing min for the RC/ATS-derived CFs, whereas at 24 min, process. For the RC/ADHP-derived CFs, irrespective a few mesopores with diameters of 20-40 nm were of the dwell time, several meso- and some macropo- found.

Figure 6: SEM pictures of CFs obtained via continuous carbonization at 1400 °C, top: RC/ADHP-derived CFs, 24 min; bottom: RC/ATS-derived CFs, 12 min.

Structural differences revealed by ADHP was 0.9 GPa. A slight increase in Young’s mo- discontinuous carbonization experiments dulus (E) of 10 GPa was reached at a carbonization In order to gain a better understanding of the differen- temperature of 1800 °C. ces between the two precursors in carbonization, structure formation of RC/ADHP- and RC/ATS-based Table 5: . Mechanical properties of RC/ADHP- and RC/ATS- CFs was studied in more detail. Precursor fibers con- derived CFs from discontinuous carbonization (average taining 1 wt.-% of P and S, respectively, were carbo- values from 20 single filament tests). nized under nitrogen in a batch furnace at different temperatures in the range 400-2200 °C. Beside of ten- Φ [μm] ε [%] σ [GPa] E [GPa] sile testing, the resulting CFs were studied via ele- RC/ADHP 10 2.4 0.9 39.0 mental analysis and WAXS. The mechanical proper- RC/ATS 10 2.4 0.9 39.6 ties of discontinuously prepared CFs did not reach the good level of mechanical properties found in conti- With increasing carbonization temperature, the C con- nuous carbonization experiments. Fibers, which were tent increased to >99 wt.-% at 1800 °C, whereas N and not treated with ADHP or ATS were even too brittle H contents decreased (Figure 7). The delta to 100 wt.-% for the testing. CFs from both precursors showed si- in the elemental analysis in Figure 7 is due to the pre- milar mechanical properties with a maximum in tena- sence of small amounts of O, P or S in the carbonized city (σ) at carbonization temperatures of 1100-1400 material. Thus, P contents of 2-3 wt.-% were found in °C (Table 5), despite their dissimilar behavior in con- RC/ADHP-derived CFs carbonized at 1100 °C, while tinuous carbonization. Via process optimization, a no S was found in RC/ATS-derived CFs carbonized at maximum tenacity of 1.3 GPa was reached with the this temperature. Furthermore, at 1400 °C, CFs derived RC/ADHP precursor in discontinuous carbonization. from RC/ATS or untreated RC (not shown) already However, the average value for RC/ATS and RC/ consisted of > 99 wt.-% C, while RC/ADHP-derived

91 LENZINGER BERICHTE 94 (2018) 85 – 94

CFs obviously contained substantial amounts of oxy- scribed before. [28-29] From 600 °C to 2200 °C, Lc gen. At higher temperatures, a release of phosphorous increased from 9 Å to 14 Å and La more than doubled and oxygen from the CF, possibly in the form of phos- from ca. 25 to 61 Å for both precursors. In the same phorous oxides or gaseous P2,[36] is indicated. Nota- temperature range, d002 decreased from ca. 4,0 Å to bly, the C contents of RC/ADHP-derived CFs were 3.6 Å, again for both precursors. This can be attributed higher after continuous carbonization than after dis- to the improvement of the graphitic structure. Up to continuous carbonization, while the overall dwell time 1400 °C, RC/ATS-derived CFs tended to smaller d002 was shorter. and larger La values than RC/ADHP-derived ones, probably because of the larger amount of heteroatoms present in the structure of RC/ADHP-derived CFs (cf. elemental analysis). The alignment of values above 1400 °C can be explained by the liberation of oxygen and phosphorous at this temperature for RC/ADHP- derived CFs. In relation to continuous carbonization, somewhat higher values for La and Lc were found at 1400 °C. This is probably due to the overall dwell time, which was longer for the discontinuous carbonization, mostly because of the long cooling time of the batch furnace. While for RC/ADHP-derived CFs the d002 value was similar for continuous and discontinuous processing, a smaller d002 value was obtained for RC/ATS-derived fibers in course of discontinuous carbonization. Again, the commercial PAN-based T300 CF, probably carbonized at a temperature of 1200- 1400 °C, was used for the comparison of structural properties. In the temperature range 1200-1400 °C, La of both RC-derived CFs reached the value of the T300 CF (La = 37.8 Å). However, Lc and d002 reached the values of the T300 CF (d002 = 3.59 Å, Lc = 1.29 nm) only at temperatures above 1800 °C. Figure 7: C, H and N contents of RC/ADHP- and RC/ATS- For all discontinuous carbonization trials the maximum derived CFs obtained via discontinuous carbonization. preferred orientation (P.O.) of the graphite planes (002) reached only 50%, even at high carbonization Figure 8 depicts the WAXS diffraction patterns of temperatures. Notably, an increase in orientation of the cellulose-based CFs at temperatures between 400 and (002) plane is a fundamental prerequisite for improv- 2200 °C and shows the growth of the crystallites and ing Young’s modulus [37-38] but requires dynamic the decrease of the interlayer spacing d002, as de- tension, as shown before.

a b

Figure 8: a) Normalized diffraction patterns of RC/ADHP-derived CFs with increasing temperature; b) development of interlayer spacing d002 and crystallite dimensions La and Lc of RC/ADHP- and RC/ATS-derived CFs with increasing carbonization temperature.

92 LENZINGER BERICHTE 94 (2018) 85 – 94

WAXS and elemental analysis revealed differences in with regard to processing conditions as well as structure the carbon structure of RC/ATS- and RC/ADHP- based and shape of the precursor fiber in order to substitute CFs. Generally, the phosphorous in RC/ADHP-based PAN-based CFs with a more sustainable cellulose- CFs impedes structure formation in the material. In the based CF. discontinuous process, a difference in La was observed rather than in Lc leading towards the conclusion that Acknowledgement for layer stacking and thus increase of Lc, also any hindrance in shrinkage is of importance. Phosphorous The authors would like to acknowledge BASF SE for and oxygen are released at sufficiently high carboniza- financing the research and support throughout the pro- tion temperature and structural differences between ject. The authors would also like to thank Cordenka RC/ATS- and RC/ADHP- based CFs diminish. How- GmbH for supply of the Rayon fibers. Furthermore, ever, such high temperatures are not favorable for the authors would like to thank Dipl.-Geol. Ulrich economic reasons and it remains unclear, whether Hageroth and Sabine Henzler for SEM and Raman orientation and structure formation at low tempera- measurements. tures could be caught up. Conclusions References

In this study, both, the carbonization yields and the [1] E. Frank, L. M. Steudle, D. Ingildeev, J. M. properties of cellulose-based CFs have been consider- Spörl, M. R. Buchmeiser, Angew. Chem. Int. Ed. ably improved without the need of an expensive high 2014, 53, 5262-5298. temperature graphitization. Standard equipment for the [2] Q.-L. Wu, S.-Y. Gu, J.-H. Gong, D. Pan, Synth. carbonization of PAN precursors can be used omitting Met. 2006, 156, 792-795. the special exhaust gas treatment which is needed in case [3] H. Zhang, L. Guo, H. Shao, X. Hu, J. Appl. of PAN-based CFs. Commercially available high- Polym. Sci. 2006, 99, 65-74. strength Rayon fibers equipped with a carbonization aid [4] J. K. Park, J. Y. Lee, Y. G. Won, D. H. Cho, were used as a precursor fiber. ATS has several advan- Method for manufacturing lyocell based carbon tages over ADHP as a carbonization aid while having fiber and lyocell based carbon fiber fabric, the identical effect on carbon yield: Dehydration US2010/0285223A1, 2010. of cellulose takes place at lower temperatures than [5] G. Bhat, K. Akato, W. Hoffman, in Proceedings with ADHP or without any additive. Concomitantly, of the Fiber Society Spring Conference, Empa carbon loss through carbon-containing volatiles and tar St. Gallen, Switzerland, 2012. is substantially reduced. Both precursor fibers were [6] X. Zhang, Y. Lu, H. Xiao, H. Peterlik, J. Mater. suitable for continuous processing with respect to Sci. 2014, 49, 673-684. their mechanical stability and resulted in CFs without [7] Sohim, http://www.sohim.by/eng/production/car- fusing of single filaments. However, RC/ATS could bon/fibers/, 2018, aufgerufen am 22.1.2018. be processed continuously over a wider range of fiber [8] N. Byrne, M. Setty, S. Blight, R. Tadros, Y. Ma, tension and led to a significant increase in textile H. Sixta, M. Hummel, Macromol. Chem. Phys. mechanical properties, especially in Young’s modulus. 2016. In case of the RC/ADHP precursor, the incorporation [9] C. E. Ford, C. V. Mitchell, Union Carbide Corpo- of phosphorous oxides retards the structure formation ration, Fibrous graphite, US3107152, 1963. which is essential for an increase in tensile strength and [10] C. B. Cross, D. R. Ecker, O. L. Stein, Union Car- modulus. The RC/ATS-derived CF show a high carbon bide Corporation, Artificial Graphite Process, content already at low temperatures and the fibers thus US3116975, 1964. withstand higher mechanical tension. Thereby, crystal- [11] P. H. Brunner, P. V. Roberts, Carbon 1980, 18, lite growth and orientation are favored. The rather ir- 217-224. regular precursor fiber shape was transferred into the [12] R. Bacon, in Chem. Phys. Carbon, Vol. 9, 1 ed. CFs. Clearly, a more defined shape and morphology (Eds.: P. L. Walker, P. A. Thrower), Marcel Dek- of the precursor should boost the mechanical proper- ker, Inc, New York [u.a.], 1973, pp. 1-102. ties of the resulting CFs. In this respect, Lyocell or [13] A. A. Konkin, in Handbook of Composites Vol. 1: IL-spun filament fibers would be promising pre- Strong Fibres, Vol. 1 (Eds.: A. Kelly, Y. N. Rabot- cursors. Structural changes induced in the CF via nov, W. Watt, B. V. Perov), Elservier Science the application of a tensile force lead to tunable Publishers B.V., Amsterdam (The Netherlands), mechanical properties. They can be further optimized 1985, pp. 275-325.

93 LENZINGER BERICHTE 94 (2018) 85 – 94

[14] I. Karacan, T. Soy, J. Mater. Sci. 2013, 48, 2009- [27] N. Byrne, J. Chen, B. Fox, J. Mater. Chem. A 2021. 2014, 2, 15758-15762. [15] R. F. Schwenker, E. Pacsu, Ind. Eng. Chem. [28] J. M. Spörl, A. Ota, S. Son, K. Massonne, F. Her- 1958, 50, 91-96. manutz, M. R. Buchmeiser, Mater. Today Com- [16] S. E. Ross, Text. Res. J. 1968, 38, 906-913. mun. 2016, 7, 1-10. [17] A. Shindo, Y. Nakanishi, I. Soma, in Appl. Polym. [29] J. M. Spörl, R. Beyer, F. Abels, T. Cwik, A. Müller, Symp., Vol. 9 (Ed.: J. Preston), Interscience Publ, F. Hermanutz, M. R. Buchmeiser, Macromol. New York [u.a.], 1969, pp. 271-284. Mater. Eng. 2017, 302, 1700195. [18] P. J. Goodhew, A. J. Clarke, J. E. Bailey, Mater. [30] P. Scherrer, Gött. Nachr. 1918, 98-100. Sci. Eng. 1975, 17, 3-30. [31] B. K. Kandola, A. R. Horrocks, D. Price, G. V. [19] W. K. Tang, W. K. Neill, J. Polym. Sci., C Polym. Coleman, J. Macromol. Sci. Polym. Rev. 1996, Symp. 1964, 6, 65-81. 36, 721-794. [20] A. Basch, M. Lewin, Text. Res. J. 1975, 45, 246- [32] S. Gaan, P. Rupper, V. Salimova, M. Heuberger, 250. S. Rabe, F. Vogel, Polym. Degrad. Stab. 2009, [21] J. B. Tomlinson, C. R. Theocharis, Carbon 1992, 94, 1125-1134. 30, 907-911. [33] R. K. Jain, K. Lal, H. L. Bhatnagar, J. Appl. Polym. [22] A. A. Morozova, Y. V. Brezhneva, Fibre Chem. Sci. 1985, 30, 897-914. 1997, 29, 31-35. [34] J. Scheirs, G. Camino, W. Tumiatti, Eur. Polym. [23] H. Li, Y. Yang, Y. Wen, L. Liu, Compos. Sci. J. 2001, 37, 933-942. Technol. 2007, 67, 2675-2682. [35] D. Radlein, J. Piskorz, D. S. Scott, J. Anal. Appl. [24] Q. Wu, N. Pan, K. Deng, D. Pan, Carbohydr. Pyr. 1991, 19, 41-63. Polym. 2008, 72, 222-228. [36] A. F. Holleman, E. Wiberg, N. Wiberg, Lehrbuch [25] Q. Liu, C. Lv, Y. Yang, F. He, L. Ling, J. Mol. der anorganischen Chemie, 102. ed., de Gruyter, Struct. 2005, 733, 193-202. Berlin; New York, 2007. [26] S. Son, K. Massonne, F. Hermanutz, J. M. Spörl, [37] F. R. Barnet, M. K. Norr, Composites 1976, 7, M. R. Buchmeiser, R. Beyer, BASF SE, Method 93-99. for production of carbon fibers from cellulose [38] D. Crawford, D. J. Johnson, J. Microsc. 1971, fibers, WO2015173243, 2015. 94, 51-62.

94 LENZINGER BERICHTE 94 (2018) 95 – 104

Continuous Alkali Treatments of Lyocell. Effects of Alkali Concentration, Temperature and Fabric Construction.

Ján Široký1,2, Avinash P. Manian2, Barbora Široká2, Richard S. Blackburn1, and Thomas Bechtold*2

1 Sustainable Materials Research Group, Centre for Technical Textiles, University of Leeds, Leeds LS2 9JT, UK 2 Research Institute of Textile Chemistry and Textile Physics, University of Innsbruck, Höchsterstraße 73, A-6850, Dornbirn, Austria. * Corresponding author: Research Institute of Textile Chemistry and Textile Physics, University of Innsbruck, Tel: +43 (0) 5572 28593, Fax.: +43 (0) 5572 28629. E-mail: [email protected].

Abstract

Continuous alkali treatment under varying NaOH concentration, temperature, and fabric construction; has a substantial influence on dimensional and mechanical properties as well as on morphological, molecular and supramolecular properties of lyocell fabrics causing changes in their structure and performance. Physico- mechanical properties and structural changes of NaOH treated fabrics were examined by dimensional changes, flexural rigidity (in dry and wet state), mass per area, water retention, physico-mechanical tests, and applying ATR-FTIR spectroscopy to evaluate changes in crystallinity indices. Maxima of TCI, LOI, shrinkage, flexural rigidity, and minima for HBI, water retention, crease recovery, tensile strength and elongation at break, were observed at concentrations of 3.33 and 4.48 mol dm-3 NaOH and at temperature of 25 °C and 40 °C. Further, it was observed that treatment temperature of 40 °C have a substantially lower impact on treated fabric properties. The cellulose II-based fabrics with different construction did not show uniform behaviour in studied properties after undergoing alkali treatment, despite they are made of the same material.

Keywords: Alkali, Fibres; Textiles; Swelling; Mechanical Properties; FT-IR; Crystallinity

Introduction Materials and Methods

The focus of the work was to examine the effect of Materials process variables and fabric construction in the treat- Plain- (1/1 weave), twill- (3/1 weave) and sateen-woven ment of lyocell with NaOH in a continuous process. (5/1 weave), desized, scoured lyocell fabrics (Tencel®, Woven lyocell fabrics of three different constructions: 140 g m-2, comprised of 50/1 Nm staple yarn) used for plain, sateen and twill, were treated with NaOH solu- this observation were kindly supplied by Lenzing AG, tions of different concentrations, at two temperatures Austria. Technical grade NaOH (ca. 50% w/w) was (25°C and 40°C), in a pilot-scale open-width washing used in formulating the alkali solutions in treatments, range, to simulate commercial processing operations. with Lyogen MC (Clariant, Basel, Switzerland) added Previously [1-3], we have reported on the influence of as wetting agent. Analytical grade acetic acid was used treatment process variables on the mechanical and in formulating the neutralization liquor. physico-chemical properties of plain-woven fabrics. In this paper, we examine the interaction effects of process variables with weave type.

95 LENZINGER BERICHTE 94 (2018) 95 – 104

Alkali treatment

Figure 1: Schematic illustration of apparatus used in continuous alkali treatment process [2].

A detailed description of the process conditions and and then laid flat on tables at ambient temperature to adjustments of continuous alkali pre-treatments are remove any residual damp. available elsewhere [2]. Briefly, the continuous process of alkali pre-treatment was conducted using an apparatus Evaluations of treated samples (see Figure 1) divided into four compartments (A, B, The dimensional change was determined from the C, D); each with two sub-compartments (1 and 2) that mass/area measured on discs of 38 mm diameter. The could be heated independently. There were four stages: water retention value was determined with the centri- alkali treatment (A2), stabilization (B2), hot water fuge filtration method as per ISO 23714:2007 on 0.5 g washing (C1 and C2), neutralisation (D1), and final hot specimens at 2790 ×g. The crease recovery angle (CRA) water washing (D2). The fabric was passed through was measured as per DIN 53890 after 30 min of the apparatus over a series of rollers including tension recovery along both warp and weft directions. The compensators (T) and pressurised squeeze rollers (P). abrasion resistance was determined as per ISO 12947- The fabric after passing through the last compartment 3:1999 from the mass loss in specimens after 7000 (D) was wound on a take-up roller (R). cycles on a Martindale abrasion tester (James H. Heal & Co. Ltd, Halifax, England). The force and elonga- The NaOH concentrations in the treatment bath (A2) tion at break along the warp direction in samples was were 2.25, 3.33, 4.48, 4.65 or 7.15 mol dm-3 and the tested with the strip method as per ISO 13934-1:1999 temperatures were set at 25 °C or 40 °C. The cor- with a 200 mm gauge length at a 20 mm min-1 rate of responding NaOH concentrations in the stabilization extension. The flexural rigidity was determined from bath (B2) were 0.5, 0.6, 0.8, 1.0 or 1.4 mol dm-3 bending lengths measured as per BS 3356:1990 on a respectively, and the temperature was set always at home-made apparatus. The specimens were then padded 60 °C. The hot water baths (C1, C2, D2) contained through deionized water at a nip pressure of 1 bar, and soft water at 80 °C, while the neutralization bath (D1) their “wet” flexural rigidity was determined. The crys- contained 2 g dm-3 acetic acid (80% v/v) at 80 °C. The tallinity changes were determined with Attenuated Total liquor volume in each sub-compartment was 20 dm3. Reflectance – Fourier Transform Infrared (ATR-FTIR) The tension compensators in A and B were set at Spectroscopy, on a Perkin-Elmer Spectrum BX spectro- 147 N m-1 of fabric width, and those in C and D at photometer with diamond ATR attachment. The sum 49 N m-1 of fabric width. The pressure roll at the end of 64 scans over 4000-600 cm-1 at a resolution of of compartment A was set to 2 bar, and all other rolls 4 cm-1 were recorded at each of four points per sample. were set to 6 bar. The speed of passage of fabric through the system was set at 2 m min-1. A control treatment was Results and discussion. performed with soft water at 25 °C in the treatment bath (A2), and soft water at 60 °C in the stabilization bath In the discussions that follow, the samples are labelled (B2), with all other treatment parameters the same. 147/25 and 147/40 to denote the tension/temperature in the alkali treatment bath. To include values from un- The treated fabrics were removed from the take-up treated and control samples in plots, they are assigned roller and dried by passage through a stenter heated to the nominal NaOH concentrations of 0.5 and 1.0 mol 130°C at a speed of 1 m/min (residence time of 60s), dm-3 respectively.

96 LENZINGER BERICHTE 94 (2018) 95 – 104

Evaluation of changes in crystallinity and hydrogen bonding intensity by ATR-FTIR

Figure 2: Total crystallinity index (TCI, blue diamond), lateral order index (LOI, red square) and hydrogen bonding intensity (HBI, green triangle) of plain (a, b), twill (c, d) and sateen (e, f) with increasing NaOH concentration, under constant tension (147 N m-1) and varying temperature (25°C for a, c and e, or 40°C for b, d and f) in treatment stage of continuous NaOH treatment.

The evaluations are based on indices, which are ratios The crystallinity changes in the lyocell fabrics sub- of absorbance intensities at two wavenumbers. These jected to alkali treatment, obtained by using ATR-FTIR include Total Crystallinity Index (TOI), Lateral Order method, are shown in Figure 2. In general, TCI and Index (LOI), and Hydrogen-Bond Intensity (HBI) LOI show parallel behaviour over the whole range of [1, 5-9]. The TCI represents the overall degree of or- alkali concentrations and treatment temperatures, alder in cellulose and is given by the ratio of absorbance though the absolute values differ. The higher magnitude at 1364 cm-1 (C–H deformation in cellulose II) to for LOI occurred due to fact that LOI represents the 2892 cm-1 (C–H stretch in cellulose II). LOI reflects ordered regions perpendicular to the chain direction, the ordered regions perpendicular to the chain direc- which is greatly influenced by chemical processing of tion, and is the ratio of the band at 1418 cm-1 (CH2 cellulose [6]. Two peaks are observed for TCI and scissoring at C(6) in cellulose II ) to 894 cm-1 (C–O–C LOI at a concentration of 3.33 mol dm-3 NaOH for valence vibration of β-glycosidic linkage). The HBI is fabrics treated at 25 °C, and at a concentration of the ratio of absorbance bands at 3336 cm-1 to 1336 cm-1, 4.48 mol dm-3 NaOH for fabrics treated at 40 °C. These and is closely related to the well ordered crystalline maxima may have been most probably influenced phase and the degree of intermolecular regularity. by several factors. The known maximum of uptake

97 LENZINGER BERICHTE 94 (2018) 95 – 104

(about 3.00 mol dm-3) and degree of swelling, which at its maximum (between 3.00 and 3.75 mol dm-3) decreases with increasing temperature of the treatment liquor [10]. Also, the rate of fibre swelling has an influence which decreases with increasing concentration of alkali, attributed to increasing viscosity and surface tension in the treatment liquor and a reduction in the rate of alkali diffusion within substrates [1]. Another influence may be an increase of fibre orientation due to applied tension to softened – fully swollen substrate at maximum swelling (crystalline conversion to less ordered structure via breaking inter- and partly intra-hydroxyl groups within cellulose unit). Therefore, the maxima of crystallinity indices were achieved at higher NaOH Figure 3: Mass per area vs. NaOH concentration in the fab- concentration as it is observed when pad-batch pro- rics: untreated plain (white circle), twill (red snow flake) or cess is applied. However, the differences in magnitude sateen (green cross); control plain (blue diamond), twill (red for TCI, LOI or HBI among used constructions revealed square) or sateen (green triangle), and fabrics treated with some trends which may have been influenced by their different alkali concentrations (2.25-7.15 mol dm-3) and weaving and thus, different accessibility to alkali. The temperatures (25°C with full markers and lines; 40°C with similar magnitude for TCI was observed when fabrics white markers and dashed lines) in treatment stage. were treated under both treatment temperatures 25 °C or 40 °C. Whilst the LOI increased in order plain ment temperature of 25 °C, when treated with 40 °C < twill < sateen as the weaving became loose at treat- rose in order plain < twill ≈ sateen. When comparing

Figure 4: Flexural rigidity of the dry-state (a-b) warp and (c-d) weft in the fabrics: untreated plain (white circle), twill (red snow flake) or sateen (green cross); control plain (blue diamond), twill (red square) or sateen (green triangle), and fabrics following the alkali-treatment (2.25-7.15 mol dm-3), tension and temperatures (a-c; with full markers and lines) 147 N m-1 / 25°C and (b- d; with white markers and dashed lines) 147 N m-1 / 40°C.

98 LENZINGER BERICHTE 94 (2018) 95 – 104

the HBI for both temperatures, the magnitudes are control with those treated with lowest NaOH concent- more or less the same for all examined constructions. ration, the 2.25 mol dm-3, only the plain construction When the comparison of TCI or LOI with HBI is did not show substantial difference. made, a “mirror effect” is observed, while TCI or LOI Dimensional change in all fabrics examined upon treat- show maxima at particular concentration, HBI exhibit ment with water (control) or NaOH treatment was re- minima. This was previously attributed to changes lated to fibre swelling, wherein the yarns increase in with bulk density as well as with symmetry of cellu- diameter by about 35% [12], and thus, a warp thread has lose II [1] due to its close relation to the well-ordered to take a longer path around the swollen weft thread system and the degree of intra- and inter-molecular [13], which, in turn, will cause contraction [2,10,14-17]. regularity of hydrogen bonds [11]. The overall effect of the swelling mechanism on fabric dimensional stability is highly dependent on the tight- Dimensional change ness of the weave [15], and, as a result, the swollen In general, the mass/area (Figure 3) increased with a rise material becomes very compact and tight, and inter-yarn of NaOH pre-treatment concentration for plain, where- space decreases [13]. In addition, it is known that the as twill and sateen levelled off between 4.48-7.15 degree of fibre swelling and its effect on fabric dimen- mol dm-3. However, substantial differences were noted sion is influenced by factors such as NaOH concentra- when comparing the various constructions: mass/area tion, temperature and mechanical restrictions imposed increased in the order plain < twill < sateen. Treatment on fibres [2,10,14-17]. Also, the mass/area indicates a temperature showed either very little or no difference. significant influence of fabric construction upon solution Increase in control substrate in all examined construc- uptake/swelling and its effect on fabric dimensional tions can be seen in comparison to the un-treated sub- stability. As the fabric weave becomes more free, plain strates, indicating a degree of overall fabric shrinkage < twill < sateen, the swelling and the mass/area rises, upon treatment with water [2]. In the comparison of and propensity to shrink increases.

Figure 5: Flexural rigidity of the wet-state (a-b) warp and (c-d) weft in the fabrics: untreated plain (white circle), twill (red snow flake) or sateen (green cross); control plain (blue diamond), twill (red square) or sateen (green triangle), and fabrics following the alkali-treatment (2.25-7.15 mol dm-3), tension and temperatures (a-c; with full markers and lines) 147 N m-1 / 25°C and (b-d; with white markers and dashed lines) 147 N m-1 / 40°C.

99 LENZINGER BERICHTE 94 (2018) 95 – 104

The obtained results from measurement of flexural shape of fibres [19, 20] as they became wedged against rigidity in all studied fabrics are shown in Figure 4 and each other within yarns and led to its close packing. Figure 5 in the dry-state and wet-state in both warp This close packing of fibres within the yarns led further and weft directions. The dry flexural rigidity, Figure 4, to increase yarn and fabric flexural rigidity. These dis- was greater in the NaOH treated fabrics compared to cussed effects may have been interconnected with the untreated or control fabrics. The control plain fabric ex- obtained peaks in flexural rigidity of the fabrics treated hibited higher value than that untreated along the warp at a concentration of 3.33 mol dm-3 NaOH for fabrics direction, but similar values were observed along the treated at 25 °C, and at a concentration of 4.48 mol weft direction. In the case of twill and sateen fabrics dm-3 NaOH for fabrics treated at 40 °C. Upon re- such a difference was not observed, the un-treated and wetting in water, the degree of yarn swelling did not control fabrics were similar in both directions. Among reach the levels attained during alkali treatment and thus, NaOH treated fabrics, the warp dry flexural rigidity inter-yarn spaces were maintained at the cross-over peak was observed at 4.48 mol dm-3 NaOH at 25 °C, points, and the wet flexural rigidity was reduced [2]. and at 4.65 mol dm-3 at 40 °C, respectively. It was also observed that two maxima occur for weft dry flexural Water retention values (WRV) rigidity at a concentration of 3.33 mol dm-3 NaOH for WRV gives a quantitative indication of the spatial samples treated at 25 °C, and at a concentration of changes in the non-crystalline regions of the fibres. 4.48 mol dm-3 NaOH for samples treated at 40 °C, apart Due to the relation between swelling properties and from the twill fabric treated at 40°C where the maximum fine structure of cellulose fibre, WRV might be an in- was at concentration of 4.65 mol dm-3. dicator of the changes in supramolecular structure Wet flexural rigidity, Figure 5, was generally lower in such as degree of crystallization, amorphous and void alkali treated fabrics, apart from the warp sateen fab- regions. WRV strongly correlates with the volume and ric treated at temperature of 25 °C, compared with the structure of pore [21], and it is inversely related to untreated or control fabrics. There were no observed crystallinity. However, WRV values include not only significant differences between untreated and control water from pores but also bound water and surface fabrics along the warp direction, but different values water as well. were detected for twill or sateen fabrics along the weft direction. Two peaks, dependent on the treatment temperature, were observed in both warp and weft wet flexural rigidity of alkali treated fabrics, at a concentration of 3.33 mol dm-3 NaOH for fabrics treated at 25 °C, and at concentration of 4.48 mol dm-3 NaOH for fabrics treated at 40 °C. Obviously, twill fabrics at lower treatment temperature showed an exception to this observation where the peak is marginally towards concentration of 4.48 mol dm-3 NaOH, however, increased error bar needs to be considered. Flexural rigidity is influenced by the physical structure of the fabrics, with therefore different interlacements or cross-over points in the fabric weave. Mobility of Figure 6: Water retention in the fabrics: untreated plain the yarns within fabrics is influenced by the compact- (white circle), twill (red snow flake) or sateen (green cross); ness of the yarns; close contact among the yarns reduces control plain (blue diamond), twill (red square) or sateen mobility and enhances the flexural rigidity, in contrast, (green triangel), and fabrics following the alkali-treatment inter-yarn spaces enhance the mobility and reduce the (2.25-7.15 mol dm-3), tension and temperatures (a-c; with full flexural rigidity. Therefore, the aqueous solution pick- markers and lines) 147 N m-1 / 25°C and (b-d; with white up differs and it is strongly influenced with swelling. markers and dashed lines) 147 N m-1 / 40°C.white markers The high dry flexural rigidities obtained herein for and dashed lines) in treatment stage. alkali treated fabrics may have been due to alkali treatment under tension which after neutralization and WRVs (Figure 6) were higher for NaOH treated fabrics drying leads generally to high dry flexural rigidity compared to untreated and control fibres for the same [18]. Those increased values resulted from the tem- construction, which is as a result of increases in capillary porary fabric settings. Also, the fabric flexural rigidity spaces caused by changes to fibre conformation within can be influenced by the rigidity of the yarns, great the yarns as well as structural reorganization in the swelling in alkali solutions affects and forms the course of swelling, de-swelling or degree of swelling.

100 LENZINGER BERICHTE 94 (2018) 95 – 104

Over the range of NaOH concentrations used, a gradual Tensile strength and elongation. decrease was detected at both temperatures (25 °C and Values of tensile strength and elongation at break are 40 °C), as well as clear order of plain < twill < sateen. also presented in Table 1. In general, twill and sateen Two minima were observed for fabrics treated at 25 °C; fabrics exhibited higher tensile strength in comparison first at 3.33 mol dm-3 NaOH for plain fabric, and second with plain fabric. Treatment temperature does not show at 4.65 mol dm-3 for twill, but sateen levelled of between significant influence on tensile strength of the same 3.33-7.15 mol dm-3. While, treatment temperature type of fabric. Compared to the control fabric, only of 40 °C led to sharp minimum at 4.48 mol dm-3 for all twill showed a distinct increase in tensile strength treated fabrics. The minimum in WRV at 4.65 mol dm-3 (3.0-11.6% for 25 °C; 0.3-10.0% for 40 °C) over the NaOH may imply a maximum swelling or reaching whole range of alkali concentrations and temperatures maximum swelling at the higher NaOH concentration in used. In contrast, sateen fabric exhibited lower values comparison with plain fabric due to fabric construction over the whole range of alkali concentrations and differences. Differences between fabric constructions temperatures used. From the alkali treated plain fab- are associated with different interlacements within each ric, only 2.25 mol dm-3 NaOH for both temperatures fabric construction and, thus, the disparate aqueous exhibited higher values in comparison with the con- NaOH uptake and kinetics apply herein [13] as well trol fabric; other alkali concentrations applied to plain as the greatest reorganization in amorphous and quasi- fabric displayed lower or the same values. crystalline structure observed previously.[1] When Opposite trends were observed for flexural rigidity, the effect of temperature is considered, the elevated wherein maxima for flexural rigidities were detected temperature exhibited lower WRV for all studied con- at the same points associated with minimum tensile structions and thus, more expressive effect of struc- strength; this is associated to fabric embrittlement ture can be expected. caused by NaOH treatment. In all examined process conditions and different constructions, the highest Abrasion resistance. elongation at break was observed in fabrics subjected From mass loss in the abrasion resistance tests shown to 7.15 mol dm-3 NaOH, with very little or no diffe- in Table 1, it was observed that the alkali treated fab- rence between treatments at 25 °C and 40 °C. Com- rics showed enhanced abrasion resistance compared pared to the control fabric, elongation at break for plain to the control fabric, and the process conditions, alkali fabric increased by 27.9%, twill by 30.0% and sateen concentration and treatment temperature exerted sig- by 18.8% (average values between two temperatures), nificant influence on the abrasion resistance. Signifi- which may have been influenced by an increase in yarn cant improvements in abrasion among NaOH treated crimp and, thus, fabric tends to elongate the most. fabrics were observed at 3.33 mol dm-3 NaOH for Overall, fabric elongation at break decreased with in- 25 °C, and at 4.48 mol dm-3 NaOH for 40 °C. How- creasing alkali concentration in the range of 2.25-4.65 ever, an exception is found for plain fabric treated with mol dm-3 NaOH; however, some differences between 3.33 mol dm-3 at 25°C where there is a marginal im- fabric construction and treatment temperature were provement in abrasion among NaOH treated fabrics, in detected. In the fabrics treated at 25 °C, elongation at fact, minimum in mass loss was observed at 7.15 mol break decreased continuously with alkali concentra- dm-3 NaOH. In general, probably as a result of the de- tion up to either 4.48 mol dm-3 NaOH for plain fabric crease of inter-yarn spaces at cross-over points in the and 3.33 mol dm-3 NaOH for twill and sateen fabrics, treated fabrics as the mass/area increased (see Figure 4), increasing thereafter. Conversely, elongation at break and therefore, treated fabrics became packed together. of fabrics subjected to 40 °C reduced gradually with This leads to a greater propensity in the treated fabrics alkali concentration up to 4.65 mol dm-3 NaOH, but for the dissipation of applied stress. sharp increase was observed in fabrics treated at a con- In addition, the abrasion resistance of the control sateen centration of 4.48 mol dm-3. The height of the particular fabric was similar to that sateen fabric treated with peaks is reduced in order plain > twill > sateen. 2.25 mol dm-3. However, with additional increase of NaOH concentration (3.33 mol dm-3), exceptional improvement in abrasion resistance is observed, 51% increase in abrasion resistance in comparison to control sateen fabric. It is believed that at this particular process conditions at which sateen fabric reaches kind of “optimum” treatment towards enhancement of this property.

101 LENZINGER BERICHTE 94 (2018) 95 – 104

Crease recovery at 25 °C, and at 4.48 mol dm-3 NaOH concentration for The alkali concentration range of 2.53-7.15 mol dm-3 fabrics treated at 40 °C. These treatment parameters, NaOH, and treatment temperature 25 °C or 40 °C, did which are required to reach maximum swelling, are not exhibit a significant influence on the crease re- markedly higher compared to pad-batch process (with- covery (shown in Table 1 as crease recovery angle; out tension). Those maxima are mainly affected by CRA) between different fabric constructions. Fabrics maximum swelling and applied tension, which tends treated at elevated temperature (40 °C) displayed sub- to increase the fibre orientation of already softened substantially greater crease recovery than those treated at strate due to alkalisation. Under the same conditions, lower temperature (25 °C). A sharp minimum at a con- the greatest molecular reorganization in the amor- centration of 3.33 mol dm-3 NaOH was observed for phous and quasi-crystalline structure of examined all examined fabrics treated at 25 °C, however, at lyocell fabrics is observed; treated lyocell showed the 40 °C a minimum at 4.48 mol dm-3 NaOH for plain maxima for TCI and LOI, and minima for HBI. This fabric was observed, and at 4.65 mol dm-3 NaOH was evidenced by dimensional and mechanical proper- for twill fabric; no distinguishing minimum for sateen ties; peaks of fabric shrinkage and flexural rigidity (in fabric was observed. Fabric crease recovery is in- dry and wet state), and minima in WRV, crease reco- versely related to the rigidity and directly related very, tensile strength and elongation at break. to the elastic recovery in fabrics [2], as was observed Degree of swelling increases as temperature decreases herein. [10,22], and has a substantial effect on examined fabrics performance such as accessibility, flexibility, softness, Conclusions and extensibility of the treated substrate in comparison to, for example, the pad-batch process. Overall, the tem- In this work, it was demonstrated that the most sub- perature of 25 °C has a significantly higher influence in stantial changes occur at certain alkali concentrations comparison with treatment at 40 °C. Therefore, the ele- and temperatures when tension of 147 N m-1 is applied; vated treatment temperature (40°C) in conjunction with at 3.33 mol dm-3 NaOH concentration for fabrics treated NaOH concentration of 4.48 mol dm-3 have a substan-

Table 1: Mechanical properties.

a Alkali concentrations [mol dm-3] in treatment / stabilization liquors. b Temperature of treatment liquor; stabilization liquor temperature was maintained at 60°C in all experiments. c Mass loss observed after 7000 abrasion cycles in tests. d Measurements along warp direction in fabrics. e CRA after 30 min of recovery; sum of values obtained along warp and weft directions.

102 LENZINGER BERICHTE 94 (2018) 95 – 104

tially less negative impact on treated fabric properties [5] Corgié, S.C., H.M. Smith, and L.P. Walker, En- and crystallinity indices, except the dimensional change, zymatic transformations of cellulose assessed by which is more pronounced. Additionally, at treatment quantitative high-throughput fourier transform in- condition of 4.48 mol dm-3 NaOH and temperature of frared spectroscopy (QHT-FTIR). Biotechnology 40 °C, the elongation showed increase between 8-10% and Bioengineering, 2011. 108(7): p. 1509-1520. for all studied constructions compared to that control [6] Hofmann, D., H.P. Fink, and B. Philipp, Lateral fabric, in other words, they became more extensible. Crystallite Size and Lattice-Distortions in Cellulo- Distinguished differences in swelling and overall per- se-II Samples of Different Origin. Polymer, 1989. formance of lyocell fabrics with different construc- 30(2): p. 237-241. tion, but consisting of the same type of cellulosic fibre [7] Nelson, M.L. and R.T. O’Connor, Relation of and of similar mass per area were observed. These Certain Infrared Bands to Cellulose Crystallinity changes can point out the significance of taking them and Crystal Lattice Type. Part II. A New Infrared on account particularly in the course of the end use, Ratio for Estimation of Crystallinity in Celluloses e.g. in fabric drape, softness or roughness, dyeing and I and II. Journal of Applied Polymer Science, finishing, etc. 1964. 8(3): p. 1325-1341. [8] Hurtubise, F. G. and H. Krassig, Classification of Acknowledgements Fine Structural Characteristics in Cellulose by In- frared Spectroscopy – Use of Potassium Bromide The authors thank The University of Leeds and Pellet Technique. Analytical Chemistry, 1960. 32(2): Lenzing AG for the provision of a scholarship to Mr. p. 177-181. Široký, and we are grateful to the Versuchsanstalt- [9] O’Connor, R.T., E.F. DuPré, and D. Mitcham, Textil and HTL-Dornbirn for the use of their facilities, Applications of Infrared Absorption Spectroscopy the Amt der Vorarlberger Landesregierung, Europäi- to Investigations of Cotton and Modified Cottons scher Fonds für Regionale Entwicklung (EFRE), and Part I: Physical and Crystalline Modifications and COMET K-project ‘‘Sports Textiles’’ project number Oxidation. Textile Research Journal, 1958. 28(5): 820494 funded by Die Österreichische Forschungs- p. 382-392. förderungsgesellschaft (FFG), for their provision of [10] Warwicker, J., Swelling of Cotton in Alkalis and the financial support. The authors would also like to Acids. Journal of Applied Polymer Science, 1969. acknowledge Dr. Mohammad Abu-Rous and Dr. Jörg 13(1): p. 41-54. Schlangen for their assistance, help with experimental [11] Široký, J., et al., Alkali treatment of cellulose II work, and also for fruitful discussion of the results. fibres and effect on dye sorption. Carbohydrate Polymers, 2011. 84(1): p. 299-307. References [12] Burrow, T., Recent results with lyocell fibres in textiles. Lenzinger Berichte, 1998. 78: p. 37-40. [1] Široký, J., et al., Attenuated total reflectance [13] Široký, J., R.S. Blackburn, and T. Bechtold, In- Fourier-transform Infrared spectroscopy analysis fluence of fabric structure on NaOH release from of crystallinity changes in lyocell following con- woven lyocell type material. Textile Research tinuous treatment with sodium hydroxide. Cellu- Journal, 2011. doi:10.1177/0040517511399968. lose, 2010. 17(1): p. 103-115. [14] Bredereck, K., et al., Alkali- und Flüssigammoniak- [2] Široký, J., et al., Alkali Treatments of Lyocell in Behandlung von Lyocellfasern. Melliand Textil- Continuous Processes. I. Effects of Temperature berichte, 2003. 84(1/2): p. 58-64. and Alkali Concentration on the Treatments of [15] Saville, B.P., Physical testing of textiles. 1999, Plain Woven Fabrics. Journal of Applied Polymer Cambridge: Woodhead Publishing. Science, 2009. 113(6): p. 3646-3655. [16] Richter, G.A., L.E. Herdle, and I.L. Gage, Acces- [3] Široký, J., et al., Continuous alkali treatment pro- sible Cellulose as Measured by Sorption of Sulfuric cess of plain woven lyocell fabrics: Effect of alkali Acid. Industrial and Engineering Chemistry, 1953. concentration. Itc&Dc: 4th International Textile 45(12): p. 2773-2779. Clothing & Design Conference, Book of Proceed- [17] Neale, S.M., The swelling of cellulose, and its ings, 2008: p. 453-458. affinity relations with aqueous solutions. Part I – [4] Široká, B., J. Široký, and T. Bechtold, Application Experiments on the behaviour of cotton cellulose of ATR-FT-IR Single-Fiber Analysis for the Iden- and regenerated cellulose in sodium hydroxide tification of a Foreign Polymer in Textile Matrix. solution, and their theoretical interpretation. Jour- International Journal of Polymer Analysis and nal of the Textile Institute Transactions, 1929. Characterization, 2011. 16(4): p. 259-268. 20(12): p. T373-T400.

103 LENZINGER BERICHTE 94 (2018) 95 – 104

[18] Ibbett, R.N. and Y.L. Hsieh, Effect of fiber swel- perties of lyocell. European Polymer Journal, ling on the structure of lyocell fabrics. Textile Re- 2009. 45(2): p. 455-465. search Journal, 2001. 71(2): p. 164-173. [21] Schaumann, W., Properties of Lenzing viscose [19] Goswami, P., et al., Dyeing behaviour of lyocell and Lenzing modal applied to finishing. Lenzin- fabric: effect of NaOH pre-treatment. Cellulose, ger Berichte, 1996. 75: p. 81-90. 2009. 16(3): p. 481-489. [22] Krässig, H. A., Cellulose: Structure, Accessibility [20] Goswami, P., et al., Effect of sodium hydroxide and Reactivity. 1993, Gordon and Breach Science pre-treatment on the optical and structural pro- Publishers S. A: Yverdon, Switzerland.

104 LENZINGER BERICHTE 94 (2018) 105 – 113

Chitin Coated Cellulosic Textiles as Natural Barrier Materials

Auriane Freyburger1, Yaqing Duan2, Cordt Zollfrank2, Thomas Röder3, and Werner Kunz1,*

1 Institute of Physical and Theoretical Chemistry, University of Regensburg, Regensburg, Germany 2 Biogenic Polymers, Technical University of Munich and Straubing Center of Science for Renewable Resources, Straubing, Germany 3 Lenzing AG, Werkstraße 2, 4860 Lenzing, Austria * Contact: [email protected]

Abstract

The two most abundant biopolymers chitin and cellulose have unique properties suitable for designing renewable materials. Cellulose already plays a significant role in the production of daily materials such as textile fibers while chitin still remains widely less utilized. In this work, new cellulose/chitin composite materials were prepared to study the advantages of a chitin coat on the properties of cellulose-based textiles. Textiles produced from TENCEL®, Lenzing Viscose®, Lenzing Modal® and cotton fibers were coated by regenerating chitin from a mixture of an ionic liquid (1-butyl-3-methylimidazolium acetate) and a bio- sourced co-solvent (γ-valerolactone). The coat was applied only on one side of the textiles. Structural and morphological analysis demonstrated that a uniform and non-modified chitin film of 10 µm was successfully coated on all type of textiles without damaging their network. The degree of acetylation of the used chitin and the fabricated chitin coating determined by elemental analysis was 95.8% and 91.9%, respectively. The influence of the chitin coat on the textiles properties was characterized by wetting, and gas/water permeability comparative studies of the conventional and coated textiles. The results demonstrated that the presence of chitin decreased the water wettability of the textiles solely on the coated site. Moreover, the chitin layer acted as a promising water and oxygen barrier independently of the nature of the textiles.

Keywords: Chitin coating, water and oxygen barrier, TENCEL®, Lenzing Viscose®, Lenzing Modal®, cotton, textile fibers.

Introduction

Currently, most polymeric materials are manufactured degradability, they have good chemical and mechanical from petrochemical (fossil) resources. In order to re- properties suitable for various applications as filtration duce the dependence on petrochemical feedstocks and process, textile, hygienic, packaging and biomedical their environmental impacts, the development and the utilization [2,4,5]. However, being not meltable below use of polymers from biomass as renewable materials their degradation temperature and having rigid bulk is important to our society [1]. In biomass, the natural structures, the processing of these biopolymers is a polysaccharides cellulose (poly β-(1,4)-D-glucopyra- real challenge and research on it has been increasing nose) and chitin (poly β-(1,4)-2-acetamido-2-desoxy- exponentially [2,6]. D-glucopyranose) are the most abundant biopolymer Cellulose processing has been successfully performed resources. With cellulose being the principal structu- by more or less complex and polluting methods and is ral component in higher plants, it has an estimated mainly industrially used in the production of fibers for annual production of 1012 tons, while chitin is mainly textile and nonwoven applications. Among the most contained in crustacean shells up to 75 000 tons per prominent man-made cellulose fiber generations one year [2,3]. In addition to their abundance and bio- finds viscose, the Modal and Tencel® fibers. The first

105 LENZINGER BERICHTE 94 (2018) 105 – 113

generation of cellulosic fibers, viscose, is manufactured Materials and Methods by derivatization of cellulose pulp with sodium hydroxide and carbon disulfide to cellulose xanthate Materials and subsequent regeneration of the cellulose by so- Textiles used in this work were obtained from Len- lution spinning. Modal is the second generation of zing AG (Austria) and produced from four different regenerated cellulose fibers and produced by a modi- cellulosic fibers: TENCEL®, cotton, Lenzing Viscose®, fied viscose process, inter alia, by using a pulp with a and Lenzing Modal®. All textiles were washed and higher degree of polymerization. The third generation desized by Lenzing AG and dried at 70 °C for 5 days named Tencel® is manufactured by the Lyocell pro- in an oven prior to use. cess, an environmentally friendlier way to produce Chitin solutions were prepared with dried α-chitin fibers from cellulose solutions in N-methylmorpholine- powder from shrimp shells purchased from Sigma N-oxide in presence of water. Besides these man-made Aldrich (Germany). Its degree of polymerization and fibers, the natural cellulosic fibers from cotton are still of acetylation were measured to be 1690 and 95.8%, the most widely applied raw materials used to produ- respectively. The solvent system used to dissolve ce textiles. [7] chitin was composed of γ-valerolactone (Reagent Chitin has not achieved the same commercial value as Plus®, purity ≥ 99%) from Sigma Aldrich (Germany) has cellulose and a large majority of its production and 1-butyl-3-methylimidazolium acetate (BmimOAc, (60-70%) is used to produce its de-acetylated deriva- purity ≥ 98%) from IoLiTec (Germany). The latter tive chitosan [8]. Chitin is structurally similar to cellu- was dried using a high vacuum setup (at 10-6 mbar) lose, but bears an acetamido group at the C2 position for 5 days, prior to dissolution of the chitin. All of the glucose units instead of an hydroxyl group. other chemicals were used without further purifi- This particularity confers it advantages over cellulose cation. such as excellent antibacterial properties, a less hydrophilic character and water retention/chelate metal ions Methods capacity [4,9]. However, it induces also a more com- Chitin solutions were prepared by dissolving α-chitin plex inter- and intramolecular hydrogen bonding net- (2 wt%) in BmimOAc (88 wt%) and GVL (10 wt%) work, causing a decreased solubility [10]. To date, the under nitrogen atmosphere at 110 °C. After total dis- only promising environmentally friendly and direct solution of chitin, a small amount of GVL was added solvents for chitin are ionic liquids (ILs). ILs are salts to dilute the highly viscous gel and to facilitate the with a melting point below 100 °C. Imidazolium-based coating. The final solution was composed of 1.6 wt% ILs, for instance 1-butyl-3-methylimidazolium acetate chitin, 30 wt% GVL and 68.4 wt% BmimOAc. (BmimOAc), were described as the most effective Chitin coating was performed by putting one face of solvents without changing drastically the properties the textiles in contact with the chitin solution at 100 °C. of chitin [10-12]. I.e. the chitin can be recovered Textiles were then moved on the chitin solution sur- without deacetylation, which is a prerequisite for main- face by a pair of tweezers for 5 minutes to provide a taining the chitin properties. homogeneous chitin layer. Chitin covered textiles The aim of this study is to give new functional proper- were left at room temperature for 2 hours to allow a ties to four different cellulosic fibers (cotton, viscose, slow gelation of the chitin. Each material was subse- modal and lyocell textiles) applying chitin coatings. quently soaked in ethanol for 2 days and in deionized For this purpose, 1-butyl-3-methylimidazolium acetate water for 2 days in order to remove the solvent resi- has been used to dissolve chitin in presence of the co- dues, i.e BmimOAc and GVL, and to induce coagula- solvent, γ-valerolactone (GVL) obtainable form re- tion of the chitin layer. The produced materials were newable resources, for facilitating the polymer solu- dried between 2 glass plates at room temperature for tion processing. By coating the chitin dope on the several days. textiles, composite materials could be prepared having Recovery of the ionic liquid BmimOAc was accom- distinguished properties of each polymer, i.e. cellu- plished after the coating procedure using the following lose and chitin, on each face. The effect of the re- method. The ethanol-based coagulation solutions were generated chitin film on the structure and the proper- first filtrated with a 0.2 μm polytetrafluorethylene ties of the textiles were characterized by scanning (PTFE) membrane filter to remove any possible poly- electron microscopy, infrared spectroscopy, water con- mer residues. The obtained liquids were dried for two tact angle measurements and water/gas permeability hours under reduced pressure (100 mbar) at 40 °C using studies. Furthermore, a method was also used to re- a rotary evaporator. Further drying was carried out cycle and reuse the costly ionic liquid at the end of the with a high vacuum setup (10-6 mbar) at 40 °C for coating route. 5 days. The purity of the recycled BmimOAc was

106 LENZINGER BERICHTE 94 (2018) 105 – 113

1 13 analyzed by H- and C-NMR measurements in the ethyl lactate and the water filled cell. The fluxJ H2O dimethyl sulfoxide-d6 recorded with a Bruker Avance can be also expressed by Equation 2: 300 spectrometer (Germany) at 300 MHz. 1 dcH2O in EL cell To characterize the composition of the materials, in- JH2O = x x V (2) A dt frared spectra were measured with a Fourier transform infrared spectroscopy (FTIR) instrument equip- with the permeation area A, the water concentration ped with an attenuated total reflectance (ATR) samp- in the ethyl lactate cell cH2O in EL cell), the volume ler by pressing the samples against a Smart Diamond related to the water concentration V, and the time t. ATR sensor (Nicolet 380, Thermo Fisher Scientific, Thus, the water permeability coefficient can be calcu- USA). Spectra were recorded in the range from 400 to lated from Equations 1 and 2 by: 4000 cm-1. V dcH2O in EL cell Films morphologies were studied by scanning electron |PH2O| = x (3) A x |∆c 2 | dt microscopy (SEM). Dried samples were placed on H O carbon tape and coated with Au/Pd. SEM images were Water permeability coefficients were calculated con- obtained with a digital scanning electron microscope sidering only the average of the values between 120 (DSM 940 A, Zeiss, Germany) in secondary imaging and 420 min. Before 120 min, the difference in water mode at an acceleration voltage of 10 kV. concentration between the two cells could not be pre- The wetting properties of the films were characterized cisely determined due to the uncertainty of the water by water contact angle measurements using a gonio- concentration measurements in the pure water cell meter type P1 equipped with a microscope and a back carried out by Karl-Fischer-Titration. light (Erna Inc., Japan). Measurements were per- Oxygen permeation under dry condition was evaluated formed at room temperature on both surfaces of the with an optical measurement conducted by a chemical coated films, i.e. on the chitin layer and the untreated optical sensor (type PSt6 from PreSens, Germany). textile layer. A 2 µL drop of deionized and sterile fil- The experimental setup consists of a self-developed trated water was formed by an automated micrometer measurement cell comprising the chemical optical pipette (Hamilton Company, USA) and placed on the sensor spot inside, read out via a polymer optical fiber sample surface. The angle of the tangent formed at the (POF-2SMA from PreSens, Germany) connected to a water drop base was recorded as well as the liquid fiber optic oxygen transmitter (Fibox 4 trace from drop absorption time on each material. Overall five PreSens, Germany). The film was fixed on the top of the drops were placed on various surface locations for permeation cell with a constant volume of 49.38 cm3 each sample. and was in contact with the ambient air through a hole The ability of water and oxygen to pass the materials of 5 mm diameter. After flushing the chamber three was investigated with permeability tests at a times with nitrogen, the increase of the oxygen partial constant temperature of 20 ± 1 °C. Water permeability pressure over time in the chamber was recorded three measurements were performed using a setup with two times for each sample. The time range was selected cells allowing the film to be in contact with 8 mL of from 0 to 13 min, which is the period of time where two different solvents, deionized water and ethyl accurate oxygen partial pressures (from 0 to 52 hPa) lactate (EL), through a hole of 3 mm in diameter. For could be measured by the sensor spot for the textiles. the coated materials, the chitin coat was in contact The permeability of the film was calculated with the with the water cell. Both solutions were stirred at 200 oxygen permeability coefficient,P O2, determined using rpm for 7 h and 150 µL were collected with an Eppen- an adaptation of the ideal gas law (4) and the Fick's dorf pipette from each cell at the same time at defined first law (5). time intervals. The water concentration in each cell in function of time was measured by volumetric Karl- pO2 x V = nO2 x R xT (4) Fischer-Titration (870 KF Titrino plus, Metrohm,

Switzerland). The water permeability coefficient,P H2O) 1 dnO2 was used to characterize the water barrier property JO2 = -PO2 x ∆cO2 = x (5) A dt of the films and was determined using Fick's first law as [13]: In Equation 4, pO2 is the partial pressure of oxygen, V

the volume of the chamber, nO2 the oxygen amount in

JH2O = PH2O × ∆cH2O (1) moles, R the gas constant, and T the absolute tempera-

ture. In Equation 5, JO2 represents the flow of oxygen per where JH2O represents the flow of water per unit area, unit area, ∆cO2 the difference in oxygen concentration

∆cH2O the difference in water concentration between between the filled chamber and the outside, and A the

107 LENZINGER BERICHTE 94 (2018) 105 – 113

permeation area. By combining these two equations, the ture composed of the ionic liquid BmimOAc and the oxygen permeability coefficient could be calculated by: co-solvent GVL, a chitin layer was formed on one side of the textiles by promoting its precipitation in etha- V dpO2 |PO2| = x (6) nol [10]. Once the textiles were coated with chitin, A x |∆P | dt O2 they exhibited a thin transparent shiny film on the chitin The oxygen permeability coefficients were calculated side while the untreated side was visually not modi- considering only the average of the calculated values fied during the coating (Figure 1). between 1 and 13 min. The morphology of the prepared materials was observed with SEM. An example of the obtained SEM Results and Discussion images for the coated material viscose performed on the surface of both sides, i.e. chitin and textile layer, Structure of the coated textiles and the cross section is presented in Figure 2. On the The studied textiles consist of an ordered structural as- one hand, it could be observed that the viscose net- sembly of cellulosic fibers (TENCEL®, Cotton, Lenzing work consisting of interlaced fibers was not damaged Viscose®, and Lenzing Modal®) having a white color. by the coating procedure. The chitin side, on the other They were produced by interlacing warp fibers (longi- hand, exhibited a homogeneous and smooth surface tudinal) and weft fibers (transversal) so that each warps without crevices or flaws. The chitin layer covered the fiber passes alternately under and over each weft pores of the textile without penetrating inside. The fibers. This structural hierarchy can be well observed cross section of the coated material displayed two in Figure 2 A). After dissolution of the chitin in a mix- distinct layers, i.e. the interlaced fibers representative

Figure 1: Appearance of untreated and chitin coated textiles.

Figure 2: SEM images of the viscose textile coated with chitin obtained from (A) the viscose side, (B) the chitin side, and (C) the cross section.

108 LENZINGER BERICHTE 94 (2018) 105 – 113

of the textile side and a thin homogeneous chitin coat. constant value self-developed for visual clarity. Firstly, The same observations were made also for all three no differences appeared comparing all spectra. Second- other coated materials. The thickness of the chitin coat ly, the specific bands characteristic to chitin such as the on all materials was estimated to be 10 ± 2 µm as in- amide I doublet band located at 1654 and 1625 cm-1, ferred from the cross section SEM images. the amide II band at 1550 cm-1 and the amide III band To confirm the presence of chitin on the coated side at 1307 cm-1 could be detected for all materials. This and to exclude degradation of the textiles, all materials means that the recovered chitin was not deacetylated. were characterized by FTIR in ATR mode. Figure 3 The other bands at 3430 and 3261 cm-1 were attributed -1 shows the FTIR-ATR spectra of the chitin side for the to the OH and NH stretching, at 3090 cm to the CH3 -1 four coated textiles. The IR curves were shifted by a asymmetric stretching, and at 2875 cm to the CH3 symmetric, CH, and CH2 stretching [14]. CH deformation and C-CH3 amide stretching vibrations were assignable to the peaks at 1428 and 1373 cm-1.The intense peaks at 1154-1010 cm-1 corresponded to C-O-C and C-O stretching. The last peak at 896 cm-1 confirmed the presence of β linkage in the molecule [14,15]. This indicated that chitin was successfully coated on the textiles and no obvious degradation occurred during the preparation.

Each starting material and its corresponding side on the coated textiles are compared by FTIR-ATR curves in Figure 4. Again, IR curves for the coated textiles Figure 3: FTIR-ATR spectra of the chitin coated side for the were shifted for a better display. The spectrum for each prepared textiles. textile side was similar to the spectrum for its non-

Figure 4: Comparative FTIR-ATR spectra of each reference textile, (A) lyocell, (B) cotton, (C) viscose, and (D) modal with its corresponding untreated side on the coated materials.

109 LENZINGER BERICHTE 94 (2018) 105 – 113

coated material independent of its crystalline structure, ting properties on one side of the coated materials, by namely cellulose I for the natural cotton and cellulose II making the textiles more hydrophobic and by slowing and amorphous cellulose for the man-made lyocell, vis- down the water penetration into the textiles, probably cose, and modal textiles. Thus, the cellulosic fibers were by plugging the holes and the reduced wetting angle not chemically degraded during the coating process. towards the more hydrophobic chitin. In addition, the absence of the chitin characteristic bands confirmed the presence of chitin solely on the Table 1: Water permeability coefficient, PH2O, with corres- coated side of the prepared material as also observed ponding standard deviations as a function of the type of ma- under SEM. terials.

-7 Evaluation of new functional properties Materials PH2O (x10 m/s) In order to evaluate the specific properties of these Lyocell 2.0 ± 0.5 new materials, a comparison of the untreated textiles Lyocell coated with chitin 0.6 ± 0.2 and the coated ones were performed regarding their Cotton 1.6 ± 0.4 wetting and permeation performance. In a first step, we studied whether the chitin coat can Cotton coated with chitin 0.5 ± 0.2 act as a water barrier, since this biopolymer is a major Viscose 1.7 ± 0.7 structural component in the exoskeleton of marine Viscose coated with chitin 0.5 ± 0.2 crustaceans and contributes to their protection [3]. Modal 2.9 ± 0.8 Therefore, wetting properties of the films were characterized using water contact angles measurements. Modal coated with chitin 0.7 ± 0.2 Typically, contact angles lower than 90° correspond to favorable wetting of the surface (termed hydrophi- To study the influence of chitin on the water diffusion lic), whereas low wettability materials (hydrophobic) through the materials, water permeability tests for all of induce contact angles higher than 90° [16]. The rela- the coated and non-coated materials were additionally tion between the water contact angles and the type of performed under identical conditions. From these textile (coated and non-coated) is illustrated in Figure experiments, water permeability coefficients were cal- 5 A). Due to the high hydrophilic behavior of cellulosic culated according to Equation 3 and are reported in fibers and the porous structure of the textiles, no water Table 1. As expected, the permeability coefficients for contact angle could be measured for the non-coated all non-coated textiles were higher than the ones for materials as the water drop was immediately absorbed. the coated materials, namely by a factor of 3.3, 3.2, Only cotton induced a contact angle of 85 ± 5° for a 3.4, and 4.1 for Lyocell, Cotton, Viscose, and Modal, few seconds (10 s), probably due its highly ordered respectively. The water transport was faster and easier structure (cellulose I). For the coated materials, the through the unmodified textiles, because of their po- water contact angle tended to be identical on almost rous and hydrophilic nature of the textiles. As shown in all chitin sides and was approximately 90° (90 ± 4° for Table 1, water permeability values for the coated tex- the coated Lyocell, 91 ± 3° for the coated Cotton, and tiles ranged from 0.5 to 0.7 x 10-7 m/s. This indicated 90 ± 4° for the coated Viscose). Only the coated modal that the chitin coat has the same water barrier proper- textile exhibited a slight smaller value of 84 ± 3°. These ties independently of the nature of the textile and slow results proved that the chitin layers exhibit a poor down the water permeation. degree of wetting towards water. After a long contact In order to explore other barrier effects of the chitin period on the surface of the chitin layer, a decrease of layer, gas (i.e. O2) permeability tests were carried out the water contact angles were detected and the absorp- under dry conditions. Figure 6 (page 112) shows the tion times were recorded (Figure 5 B). A complete ab- evolution of the oxygen partial pressure, pO2, in the sorption of a 2 µL drop on all chitin layers was reached measurement cell as a function of time investigated for after ca. 20 min (24 ± 3 min for the coated Lyocell, all materials. Oxygen permeability coefficients were 22 ± 6 min for the coated Cotton, 21 ± 1 min for the calculated according to Equation 6 considering the quasi coated Viscose, and 19 ± 4 min for the coated Modal). linear increase of the oxygen partial pressure and are

Accordingly, the formed chitin layer is not impermeable reported in Table 2. For non-coated textiles, pO2 identi- to water but significantly retards its infiltration. cally increased from 0 to 51 hPa in 13 minutes. Further-

Since the textile side of the coated materials immediately more, it has to be noticed that this pO2 raise followed the absorbed the water drop textile wettability properties same trend as in the case when no membrane was fixed were not changed during the process (Figure 5 B). on the top of the permeation chamber. Independently Consequently, the chitin coat affected only the wet- of the textile nature, all are highly permeable to oxygen

110 LENZINGER BERICHTE 94 (2018) 105 – 113

A B

Figure 5: (A) Contact angle and (B) absorption time of a 2 µL water drop on the surface of the textile materials non-coated and coated with chitin (CS stands for Chitin Side and TS for Textile Side). which can be explained by the presence of open pores. with cellulose instead of chitin. To achieve this, a Lyo- This generated high permeability coefficients for these cell textile was coated with microcrystalline cellulose materials in the range from 9.0 to 9.1 x10-4 m/s (Table 2). under the same conditions as previously described. In contrast, the oxygen diffusion through the coated Unfortunately, it was not possible to reproduce a com- materials was much slower as shown in Figure 6 B). parably thin 10 ± 2 µm cellulose layer. Therefore a

After 13 minutes, pO2 values of 3.1, 3.0, 2.8, and 2.7 quantitative comparison could not be done, especial- hPa were recorded for the coated Lyocell, Cotton, Vis- ly for the permeability tests. However, it can be men- cose, and Modal, respectively. As a result, coated tex- tioned that cellulose coating induced a lower water tiles show significant lower permeability coefficients, contact angle of 54° instead of 90° for the chitin coa- which were ranged from 0.40 to 0.46 x10-4 m/s. Com- ting. The water drop was also two times faster absor- paring these values to those obtained for the non-coa- bed by the cellulose layer than by the chitin one. Con- ted textiles, it can be concluded that oxygen respec- cerning the gas/water permeability properties of the tively permeated through Lyocell, Cotton, Viscose, and cellulose coated Lyocell, it seemed that cellulose coa- Modal 19.8, 19.8, 21.4, and 22.8 times faster than ting was more permeable to water but slightly less to through its corresponding coated materials. oxygen than chitin coating. Further work has to be done to confirm these observations.

Table 2: Oxygen permeability coefficients, OP 2 ), with corres- At the end of the coating process, the recycling of the ponding standard deviations as a function of the type of ionic liquid was attempted for sustainability and pro- materials. fitability interests. Once chitin regenerated, the ethanol- based precipitation solutions were collected and sub- -4 Materials PO2 (x10 m/s) mitted to a drying procedure detailed in the methods. Without membrane 9.4 ± 0.4 Ethanol could be successfully removed under reduced Lyocell 9.1 ± 0.3 pressure using a rotary evaporator. However, GVL Lyocell with chitin 0.46 ± 0.02 was still present in the residue due to its remarkable low vapor pressure (0.65 kPa at 25 °C and 3.5 kPa at Cotton 9.1 ± 0.4 80 °C) [17]. A second step was thus carried out to re- Cotton with chitin 0.46 ± 0.02 move GVL by using a high vacuum setup. Analysis of Viscose 9.0 ± 0.3 the obtained extract by 1H and 13C NMR revealed that Viscose with chitin 0.42 ± 0.01 BmimOAc remained unaltered and no trace of ethanol Modal 9.1 ± 0.3 or GVL was detected. The recovered IL was reused to dissolve successfully 2 wt% of chitin as described above Modal with chitin 0.40 ± 0.01 in the dissolution experiment. Again, BmimOAc was then a second time recovered and reused without appre- To prove that these new functional properties were ciable decrease of its efficiency. caused by chitin and not just by the plugging of the textiles pores, the same coating procedure was performed

111 LENZINGER BERICHTE 94 (2018) 105 – 113

Figure 6: Oxygen partial pressure, pO2, over time measured in the O2 permeation chamber for (A) all the textile materials non- coated and coated with chitin and for (B) specific to the coated materials.

112 LENZINGER BERICHTE 94 (2018) 105 – 113

Conclusions [6] Rinaudo, M.: Chitin and chitosan: Properties and applications. Progress in Polymer Science 2006, Textiles made from TENCEL®, Lenzing Viscose®, 31, 603-632. Lenzing Modal® and cotton fibers were coated with [7] Shen, L.; Patel, M. K.: Life cycle assessment of chitin, a natural and antibacterial polymer, to produce man-made cellulose fibers. Lenzinger Ber. 2010, new functional and renewable materials. The approach 88, 1-59. used therefore was the chitin regeneration on one sur- [8] Kurita, K.: Chitin and Chitosan: Functional Bio- face of the textiles. Chitin was dissolved in the ionic polymers from Marine Crustaceans. Mar. Bio- liquid 1-butyl-3methylimidazolium acetate with a bio- technol. 2006, 8, 203-226. sourced solvent γ-valerolactone and precipitated with [9] Crini, G.; Guibal, É.; Morcellet, M.; Torri, G.; ethanol, after being in contact with the starting mate- Badot, P.-M.: Chitine et chitosane. Préparation, rials. Transparent chitin coats of 10 µm thickness were propriétés et principales applications. In Chitine et successfully created on the textiles without penetra- chitosane, du biopolymère à l'application; Presses tion into the textiles. These new composite materials universitaires de Franche-Comté, 2009; pp 19-53. exhibited the unique properties of chitin on one side [10] Wu, Y.; Sasaki, T.; Irie, S.; Sakurai, K.: A novel without degrading cellulose properties on the other side. biomass-ionic liquid platform for the utilization The chitin coat influenced the properties of the resulting of native chitin. Polymer 2008, 49, 2321-2327. composite materials by making them more hydro- [11] Qin, Y.; Lu, X.; Sun, N.; Rogers, R. D.: Dissolu- phobic and by slowing down the penetration of water tion or extraction of crustacean shells using ionic or O2 into the membrane. Further optimization, as for liquids to obtain high molecular weight purified instance the increase of the thickness of the coat, could chitin and direct production of chitin films and improve the barrier properties of these new materials, fibers. Green Chem. 2010, 12, 968-971. making them potential candidates for various applica- [12] Jaworska, M. M.; Kozlecki, T.; Gorak, A.: Re- tions as impermeable textiles for hygiene products. view of the application of ionic liquids as solvents Finally, a procedure to recover the ionic liquid during for chitin. J. Polym. Eng. 2012, 32, 67-69. the process was proposed for sustainable concerns. [13] Evans, D. F.; Wennerström, H.: 6. Bilayer systems. In The Colloidal Domain: Where Physics, Acknowledgements Chemistry, Biology, and Technology Meet; Wiley, 1999; pp 328-331. As part of the project ForCycle, this work was support- [14] Pearson, F. G.; Marchessault, R. H.; Liang, C. Y.: ed by the Bavarian State Ministry of the Environment Infrared spectra of crystalline polysaccharides. and Consumer Protection. V. Chitin. J. Polym. Sci. 1960, 43, 101-16. [15] Muzzarelli, R. A. A.; Morganti, P.; Morganti, G.; References Palombo, P.; Palombo, M.; Biagini, G.; Mattioli Belmonte, M.; Giantomassi, F.; Orlandi, F.; Muz- [1] Mohanty, A. K.; Misra, M.; Drzal, L. T.: Sustain- zarelli, C.: Chitin nanofibrils/chitosan glycolate able Bio-Composites from Renewable Resources: composites as wound medicaments. Carbohydr. Opportunities and Challenges in the Green Mate- Polym. 2007, 70, 274-284. rials World. Journal of Polymers and the Environ- [16] Yuan, Y.; Lee, T. R.: Contact Angle and Wetting ment 2002, 10, 19-26. Properties. In Surface Science Techniques; Sprin- [2] Klemm, D.; Heublein, B.; Fink, H.-P.; Bohn, A.: ger Berlin Heidelberg, 2013; pp 3-34. Cellulose: Fascinating biopolymer and sustaina- [17] Horvath, I. T.; Mehdi, H.; Fabos, V.; Boda, L.; ble raw material. Angew. Chem., Int. Ed. 2005, Mika, L. T.: γ-Valerolactone – a sustainable liquid 44, 3358-3393. for energy and carbon-based chemicals. Green [3] Khor, E.: Chapter 5: the sources and production of Chem. 2008, 10, 238-242. chitin. In Chitin: fulfilling a biomaterials promise; Elsevier, 2014; pp 63-72. [4] Ravi Kumar, M. N. V.: A review of chitin and chitosan applications. Reactive and Functional Polymers 2000, 46, 1-27. [5] Islam, S.; Bhuiyan, M. A. R.; Islam, M. N.: Chitin and Chitosan: Structure, Properties and Applica- tions in Biomedical Engineering. J. Polym. En- viron. 2016, Ahead of Print.

113 LENZINGER BERICHTE 94 (2018) 115 – 122

Improvement of Dyeing Performance of Cellulose Fibers by Pre-Treatment with Amino Cellulose

Thomas Heinze1,* Thomas Wellhöfer1, Kerstin Jedvert1,2, Andreas Koschella1, and Hendryk Würfel1

1 Institute for Organic Chemistry and Macromolecular Chemistry, Friedrich Schiller University of Jena, Humboldtstraße 10, 07743 Jena, Germany 2 Dr. Kerstin Jedvert, Bio-based Fibres, Swerea IVF, P.O. Box 104, SE-431 22 Mölndal, Sweden * Contact: Thomas [email protected]

Abstract

Dyeing experiments with differently shaped cellulose fibers (paper samples and pulp fibers) were performed before and after treatment of the samples with amino cellulose in aqueous solution. Surface coating was achieved by simple dipping. A colorimetric assay applying 4,4’-dimethoxytriphenylchloromethane/ perchloric acid was used to determine the amino group content on the paper surface, while the surface loading of pulp fibers was measured by fluorescence spectrophotometry. The amino cellulose concentration in the treatment solution can be kept as low as 0.05%. Pulp fibers were coated uniformly. Coating of paper with amino cellulose increased the dyeing performance significantly independent of the amino cellulose type used.

Keywords: adsorption; cellulose and other wood products; dyes/pigments; polysaccharides; textiles

Introduction

Dyeing of cellulose fibers in paper and textiles is surface of cellulose [5-7]. Application of cationic strongly connected with the retention of the dye mole- cellulose derivatives for modification of pulp fibers cules at the fiber surface [1]. With respect to the demand has shown to be promising to improve properties of of environmentally benign processes, it is desirable to the fiber surface (e.g. wetting and bacteriostatic be- improve the interaction between the cellulose surface havior), while preserving the inherent properties of and the dye. In textile dyeing, addition of quaternary the fibers, such as high mechanical strength and low ammonium salts has been found to improve the per- density [8]. formance of anionic dyes [2]. However, adding vari- A major advantage of the treatment using amino cel- ous compounds to the dyeing mixture, which might lulose is its simple application, being added from an improve the dyeing performance, does not reduce the aqueous solution. The main part of the cationic cellu- waste. The effluents must be treated adequately before lose remains in the filtrate and can be reused easily. they can be disposed of [3]. The cationic polysaccharide layer reverses the surface Surface modification prior to the dyeing step is a prom- charge from negative to positive and, hence, increases ising route to increase the interaction between cellulose the interaction with negatively charged dye molecules fiber and dye molecules. 6-Deoxy-6-(ω-aminoalkyl) [9,10]. Therefore, it is expected that not particularly amino celluloses have been identified as a class of bio- reactive dyes can be applied to cellulose fibers and the polymer derivatives that strongly bind to a wide variety amount of dyestuff can be reduced. of materials, including metals, glass, and various syn- The accurate determination of surface loading, i.e., the thetic polymers, forming stable layers and even mono- content of amino groups on the surface is of crucial im- layers on the surface [4]. Moreover, cationic cellulose portance for the evaluation of the dyeing performance. derivatives possess a high affinity to the anionic Several methods have been described, which are

115 LENZINGER BERICHTE 94 (2018) 115 – 122

based on a colorimetric test. Gaur et al. developed a The dyes (see Table A1 in annex), except crystal vio- method based on 4,4’-dimethoxytriphenylchloro- let and thymol blue (Acros Organics), were purchased methane [11-13]. The reagent reacts with the amino from Sigma Aldrich and used as received for the pre- groups and the excess of reagent is removed. Sub- paration of 0.1% solution. sequently, perchloric acid treatment liberates the The following papers were included in the studies: 4,4’-dimethoxytriphenylmethyl cation, whose content Kymi-Acacia (coated, paper A), Chanshu-Acacia can be colorimetrically determined at 498 nm. Primary- (coated, paper B), Nordland-Betula (uncoated, paper and secondary amines can be identified with nitro- C), Nordland-Acacia (uncoated, paper D), Yes-Silver phenylisothiocyanate-O-trityl, which is subsequently multifunction (paper E). cleaved by trifluoroacetic acid and determined by colori- metry [14]. The anionic dye Acid Orange 7 (sodium- Measurements 4-[(2-hydroxy-1-naphthyl)azo]benzenesulfonate) UV/Vis-spectra were measured with a Lambda 16 binds at amino groups at a pH value of 3. Excess of UV/Vis-spectrophotometer (Perkin Elmer) in reflect- dye is removed by washings and the adsorbed dye is ance mode. desorbed with sodium hydroxide before it is colori- Fluorescence measurements were performed with an metrically determined at a wavelength of 485 nm LS 50B fluorescence spectrometer (Perkin Elmer). [15,16]. Ninhydrin can be used for the quantification of amino groups as well. The colorless reagent reacts Methods with amino groups under formation of Ruhemann’s Adsorption of Amino Cellulose on Pulp Samples purple, whose concentration is determined by colori- Amino cellulose marked with 9-(methylaminomethyl)- metry at 570 nm [17-19]. Due to the low molar extinc- anthracene was dissolved in distilled water with an tion coefficient of ε = 15,000 M-1cm-1, a large surface adjusted pH value of ~ 3.3 (adjusted with 0.1 N HCl). or a high loading is required to obtain reliable results. The solution was diluted to a final concentration of In the present paper, we wish to report on the ability 0.01%, (w/w). Adsorption experiments were conducted of amino cellulose derivatives to improve the dyeing by adding different amounts (volumes) of the solution performance of different kinds of cellulosic materials to the same amount of pulp. The treatment was carried by a simple coating treatment with aqueous amino out on 50 g (dry mass) of pulp for each sample for cellulose solutions. 20 min at room temperature under vigorous stirring. Subsequently, the solution was removed with a syringe Experimental and was filtered through a poly(tetrafluoroethylene) filter (0.45 μm, diameter 30 mm, Genetec), and Materials measured by fluorescence spectroscopy. The fluores- 6-Deoxy-6-(2-aminoethyl)amino) cellulose (EDA-cel- cence quantum yields of the labeled amino cellulose lulose, degree of substitution of EDA groups, DSEDA, derivatives were measured for samples in water-citric 0.83 [20] and 6-deoxy-6-(2-(bis(2-aminoethyl)amino- acid and diluted hydrochloric acid both at pH value ethyl)amino) cellulose (TAEA cellulose, DS of TAEA of 3.1. Anthracene and 9,10-diphenylanthracene in groups, DSTAEA 0.60 [21] were prepared according to ethanol have been used as calibration standards (see previously published methods by conversion of cellu- annex). lose p-toluenesulfonic acid ester with the corresponding amines. Dyeing of Paper Samples For adsorption experiments on pulp fibers, two dis- Dye solutions (0.1%, w/v) were prepared according solving grade pulps were used: a pre-hydrolysis kraft to Table A1 and placed in test tubes. Paper (1-2 cm2) pulp “Purple grade” from Södra (P1) and a dissolving samples were added and treated for 10 min under grade sulfite pulp from Domsjö (P2). The carboxyl shaking. After decanting the supernatant, the wet pa- groups present in the pulps were transformed into the per was washed with 2 mL of the solvent mixture used sodium salt form according to a method published for the particular dye until no further dye is extracted. elsewhere [22]. The pulps were treated with an EDA- The wet paper was carefully removed and air-dried. cellulose marked with 9-(methylaminomethyl)anthracene for fluorescence measurements. The labelled Determination of the Amino Group Content with EDA-cellulose was prepared by simultaneous con- 4,4’-Dimethoxytriphenylchloromethane version of cellulose p-toluenesulfonic acid ester with Reagent A: 1.694 g 4,4’-dimethoxytriphenylchloro- excess ethylene diamine according to [23] and the flu- methane, 1.522 g tetrabutylammonium nitrate, and orescence dye (FD) 9-(methylaminomethyl)anthra- 0.909 g sym-collidine were dissolved in 10 mL dry cene (DSEDA 0.69, DSFD 0.08). N,N-dimethylformamide (DMF) under exclusion of light.

116 LENZINGER BERICHTE 94 (2018) 115 – 122

Reagent B: 26 mL Perchloric acid were mixed with 23 Table 1: Amount of fluorescently labelled 6-deoxy-6-(2-ami- mL methanol. noethyl)amino (EDA) cellulose (AC) adsorbed on pulp fibers The sample was covered with reagent A (2 mL) and (treatment at room temperature for 20 min). kept for 35 min at room temperature under exclusion of light. The liquid phase was removed and the sample was Volume of Amount adsorbed EDA celluloseb) of AC on fibers [g] thoroughly rinsed with anhydrous DMF and anhydrous Sample solution added (average of three methanol under exclusion of moisture. Washing was no. Pulpa) [mL] measurements) continued until filtrates remained colorless upon ad- 1 P1 2 2.08 10-4 -4 dition of reagent B. After being dried, the sample was 2 P1 5 4.70 10 3 10 9.32 10-4 covered with reagent B and shortly shaken. The absorb- P1 4 P1 30 2.95 10-4 ance at 498 nm against reagent B as blind value was -4 5 P1 50 5.00 10 measured after 2 min. Calculation of the amino group 6 P2 2 2.07 10-4 content was conducted according to equation 1. 7 P2 5 4.66 10-4 8 P2 10 9.28 10-4 9 P2 30 2.96 10-4 10 P2 50 5.00 10-4 a) P1: pre-hydrolysis kraft pulp “Purple grade” from Södra, Results and Discussion P2: dissolving grade sulfite pulp from Domsjö b) Degree of substitution of 6-deoxy-6-(2-aminoethyl) amino It was found that amino cellulose interacts very inten- groups DSEDA 0.69, DS of 6-deoxy-6-(9-(methylamino- sively with a variety of materials [4]. In the present methyl)anthracene groups DSFD 0.08 work, dyeing experiments of cellulose (paper, pulp fibers) were carried out without and with a prior treat- It can be noticed that the fibers are uniformly coated ment of the samples applying amino cellulose by dip- with the fluorescent AC derivative. The different fluo- ping or spraying. The surface loading regarding the rescence intensity reflects the different amounts of AC amino cellulose was determined colorimetrically with absorbed on the fiber surface. 4,4’-dimethoxytriphenylchloromethane. Amino Cellulose Coating of Paper Samples Adsorption of Amino Cellulose (AC) on Pulp Samples The experiments were carried out with a usual copy In a first set of experiments, pulp samples were treated paper. Due to the chemical and physical nature of pa- with an aqueous solution of EDA cellulose labeled with per as well as the high porosity, high load capacities 9-(methylaminomethyl)anthracene. The amount of poly- should be expected compared to other material like, mer adsorbed was calculated by fluorescence spectro- e.g., glass. However, it turned out that the load capa- scopy of the remaining solutions, i.e., after the treatment city with respect to the coated surface is difficult to of the pulp was carried out. There is no difference in the measure because the total surface of the paper acces- amount of AC absorbed for the different pulps, as can be sible for AC cannot be measured and, moreover, swel- seen in Table 1. It could also be observed that the AC ling processes during both coating and dyeing may is attached to the pulp fibers by radiation of the samp- influence the surface area. Thus, experiments were les with UV-light after the treatment (Figure 1). performed with uniform samples of constant dimensions. The values summarized in Table 2 shows that a small concentration of the solution of amino cellulose is needed for the coating process.

Table 2: Surface loading of paper (copy paper: Future Mul- titech, white paper for inkjet, copier, laser, fax, 75 g/m2) with aqueous solution of 6-deoxy-6-(2-aminoethyl)amino cellulose (EDA) cellulose of different concentration.

Concentration of EDA cellulose (%) Figure 1: Photography of pulp fibers radiated with UV-light 0.025 0.050 0.100 2 after treatment with amino cellulose marked with 9-(methyl- Sample surface (cm ) 2 2 2 aminomethyl)anthracene (samples 1, left and 5, right, see Absorption (diluted to 1:100) 1.017 1.162 1.169 2 Table 1). Load capacity (mmol/cm- ) 1.452 1.660 1.670

117 LENZINGER BERICHTE 94 (2018) 115 – 122

It became obvious that the concentration of EDA cellu- Dyeing of Paper lose dissolved in water as low as 0.05% is sufficient to Dyes of different chemical structure were prepared as saturate the paper surface with amino cellulose (about 0.1% (w/v) solutions (see Table A1). The dye absorp- 1.6 mmol/cm-2). The slight difference of load capacity tion was evaluated by visual inspection and by UV/ between a concentration of EDA cellulose dissolved Vis-spectrometry. The results summarized in Table 3 in water of 0.050% and 0.100% is within the error clearly indicate that surface coating with amino cellu- range of the method applied. While the coating is in- lose improves the dyeing performance. Crystal violet, complete at concentration below 0.025%, an excess of methylene blue, and methyl red are absorbed, while amino cellulose is rinsed away when applying higher the paper treated with Ruhemann’s purple does not concentrations of AC. show any color. Treatment of the paper samples with An important question is the reusability of the coating EDA leads to an increase of the color intensity in most solution. As can be seen in Figure 2, the surface load cases. It became obvious that the color intensity is capacity is almost constant and, hence, independent of distinctly influenced by the chemical structure of the the number of cycle. Therefore, the amino cellulose dye. Anionic dyes (methyl red, Ruhemann’s purple) solution can be used repeatedly without loss of activity exhibited a much better binding to the amino cellulose at least for ten times. coated surface compared with cationic dyes (crystal violet, methylene blue). It is known that cationic cellulose derivatives possess a high affinity to the negatively charged cellulose surface [6]. The surface is, therefore, charged with cationic moieties, which act as anchor for the anionic dye molecules. A similar observation was reported in the case of amino group containing nanocrystalline cellulose (NCC) [24]. They found that NCC functionalized with amino moieties could be used as an adsorbent to remove anionic dyes Figure 2: Dependency of surface loading capacity of paper by in aqueous solution. The highest efficiency was found repeated treatment using the same amino cellulose solution. under acidic conditions.

Table 3: Results of dyeing experiments of paper before and after treatment with amino cellulose derivatives.

Dye Crystal violet Methylene blue Methyl red Ruhemann’s purple a) solution 70% H2O, 30% MeOH 70% H2O, 30% MeOH 70% H2O, 30% MeOH 50% H2O, 50% EtOH

Untreated

Coated with EDA- cellulose

Dye Coomassie brilliant blue G250 Congo red Thymol blue Eriochrome black T

70% H2O, 20% MeOH, solution 50% H2O, 50% EtOH 100% EtOH 100% H2O 10% HAc

Untreated

Coated with EDA- cellulose

Coated with TAEA- cellulose a) 6-Deoxy-6-(2-aminoethyl)amino) cellulose (EDA-cellulose); 6-deoxy-6-(2-(bis(2-aminoethyl)aminoethyl)amino) cellulose (TAEA cellulose)

Coomassie brilliant blue, Congo red, and Eriochrome cellulose led to comparable dyeing efficiency, i.e., the black T absorb weakly onto the cellulose fibers, Thy- color intensity was remarkably increased and com- mol blue did not lead to a visible coloration. Those parable independent of the type of AC. It may be con- dyes were selected for more detailed investigations. cluded that the number of available amino groups Treatment of cellulose with both EDA- and TAEA does not increase the coloration. Larger differences

118 LENZINGER BERICHTE 94 (2018) 115 – 122

were observed in the case of dyes without carboxylate regime given by the manufacturer. Both Bezaktiv HE moieties. Obviously, the structural features in Thymol dyes (red and marine) were applied as 0.1% (w/v) solu- blue (hydroxyl groups) and Eriochrome black T tions solely or mixed with marine and red color. (hydroxyl- and sulfonate groups) induce a weaker Comparably weak dyeing could be observed in case interaction with the amino group-containing surface. of all dye combinations if no previous AC treatment While Thymol blue colored paper exhibited a different was carried out (Table 4). Almost no dyeing could be color intensity, different colors were observed in case observed in case of paper A. AC treatment of the of Erichrome black T. paper prior to the dyeing process increased the color The dyeing performance of commercially available intensity remarkably. Obviously, the dyes interact textile dyestuffs (reactive dye Benzaktiv HE red and much better with the amino groups compared to the marine) on office papers depending on the AC treat- hydroxyl groups. No difference between EDA- and ment was also studied in order to check if treatment TAEA cellulose could be found. Combination of both with AC leads to an increased dyeing effect. The dyeing dyes led to mixed colors, which were homogeneously process was carried out under omission of the dyeing distributed.

Table 4: Dyeing of paper samples with reactive dyes with and without amino cellulose treatment.

Dye ratio (red:marine) / paper a) Untreated Coated with EDAb)-Cellulose Coated with TAEAb)-Cellulose 1:0 Paper E

2:1 Paper A

1:1 Paper C

1:2 Paper D

0:1 Paper B a) Paper A (Kymi-Acacia, coated), paper B (Chanshu-Acacia, coated), paper C (Nordland-Betula, uncoated), paper D (Nordland- Acacia, uncoated), paper E (Yes-Silver Multifunction) b) 6-Deoxy-6-(2-aminoethyl)amino) cellulose (EDA-cellulose); 6-deoxy-6-(2-(bis(2-aminoethyl)aminoethyl)amino) cellulose (TAEA cellulose)

Conclusions References

It was shown that amino cellulose can be coated on [1] A. Stanislawska, E. Drzewinska, Retention of dyes different cellulose surfaces by a simple dipping tech- on cellulose fibres during paper production. Ann. nique. A much diluted solution of AC (0.05%) can be Warsaw Univ. Life Sci.-SGGW. For. Wood Tech- used several times without losing activity. Amino cellu- nol. 62, 256-259 (2007). lose is absorbed on the surface of cellulose fibers and [2] S. Sharif, S. Ahmad, M. M. Izhar-ul-Haq, Role turns the surface charge positive. Therefore, anionic of quaternary ammonium salts in improving the dyes bind better to the positively charged surface and fastness properties of anionic dyes on cellulose the dyeing performance can be improved. Further fibres. Color. Technol. 123, 8-17 (2007). studies should be performed to apply this approach to [3] T. Bechtold, E. Burtscher, Y.-T. Hung, Treatment textiles. In this regard it is interesting to examine the of textile wastes. Handbook of Industrial and resistance of this coating against washing. Hazardous Waste Treatment, Marcel Dekker, 379-413 (2004). Acknowledgements [4] T. Heinze, M. Siebert, P. Berlin, A. Koschella, Biofunctional materials based on amino cellulose KJ: The Swedish Research Council Formas is grate- derivatives – A Nanobiotechnological Concept. fully acknowledged for financial support. Macromol. Biosci. 16, 10-42 (2016).

119 LENZINGER BERICHTE 94 (2018) 115 – 122

[5] T. Genco, L. F. Zemljič, M. Bračič, K. Stana- [15] S. G. Hu, C. H. Jou, M. C. Yang, Surface grafting Kleinschek, T. Heinze, Characterization of vis- of polyester fiber with chitosan and the antibac- cose fibers modified with 6-deoxy-6-amino cellu- terial activity of pathogenic bacteria. J. Appl. lose sulfate. Cellulose 19, 2057-2067 (2012). Polym. Sci. 86, 2977-2983 (2002). [6] O. Grigoray, H. Wondraczek, A. Pfeifer, P. Fardim, [16] M. Pohl, N. Michaelis, F. Meister, T, Heinze, T. Heinze, Fluorescent multifunctional poly- Biofunctional surfaces based on dendronized cel- saccharides for sustainable supramolecular func- lulose. Biomacromolecules 10, 382-389 (2009). tionalization of fibers in water. ACS Sustainable [17] T. Finlay, V. Troll, M. Levy, A. Johnson, L. Hod- Chem. Eng. 5, 1794-1803 (2017). gins, New methods for the preparation of biospe- [7] K. Jedvert, T. Elschner, T. Heinze, Adsorption cific adsorbents and immobilized enzymes utilizing studies of amino cellulose on cellulosics. Macro- trichloro-s-triazine. Anal. Biochem. 87, 77-90 mol. Mater. Eng. 302, 1700022 (2017). (1978). [8] B. Vega, H. Wondraczek, C. S. P. Zarth, E. Heikkilä, [18] S. Moore, W. H. Stein, A modified ninhydrin P. Fardim, T, Heinze, Charge-directed fiber sur- reagent for the photometric determination of face modification by molecular assemblies of amino acids and related compounds. J. Biol. functional polysaccharides. Langmuir 29, Chem. 211, 907-913 (1954). 13388-13395 (2013). [19] V. K. Sarin, S. B. Kent, J. P. Tam, R. B. Merri- [9] S. Bratskaya, S. Schwarz, G. Petzold, T. Liebert, field, Quantitative monitoring of solid-phase pep- T. Heinze, Cationic starches of high degree of tide synthesis by the ninhydrin reaction. Anal. functionalization: 12. Modification of cellulose Biochem. 117, 147-157 (1981). fibers toward high filler technology in papermak- [20] M. Zieger, M. Wurlitzer, C. Wiegand, K. Redder- ing. Ind. Eng. Chem. Res. 45, 7374-7379 (2006). sen, S. Finger, P. Elsner, P. Laudeley, T. Liebert, [10] I. Szilagyi, G. Trefalt, A. Tiraferri, P. Maroni, T. Heinze, U.-C. Hipler, 6-Deoxy-6-aminoethy- M. Borkovec, Polyelectrolyte adsorption, inter- leneamino cellulose: synthesis and study of hemo- particle forces, and colloidal aggregation. Soft compatibility. J. Biomater. Sci., Polym. Ed. 26, Matter 10, 2479-2502 (2014). 931-946 (2015). [11] R. Gaur, S. Paliwal, P. Sharma, K. Gupta, A sim- [21] L. C. Fidale, M. Nikolajski, T. Rudolph, S. Dutz, @ ple and sensitive spectrophotometric method for F. H. Schacher, T. Heinze, Hybrid Fe3O4 amino the quantitative determination of solid supported cellulose nanoparticles in organic media – Hetero- amino groups. J. Biochem. Biophys. Meth. 18, geneous ligands for atom transfer radical poly- 323-329 (1989). merizations. J. Colloid Interface Sci. 390, 25-33 [12] R. K. Gaur, K. C. Gupta, A spectrophotometric (2013). method for the estimation of amino groups on [22] T. Köhnke, H. Brelid, G. Westman, Adsorption of polymer supports. Anal. Biochem. 180, 253-258 cationized barley husk xylan on kraft pulp fibres: (1989). Influence of degree of cationization on adsorption [13] R. K. Gaur, P. Sharma, K. C. Gupta, 4, 4’-Di- characteristics. Cellulose 16, 1109-1121 (2009). methyloxytrityl chloride: a reagent for the spectro- [23] T. Heinze, A. Pfeifer, A. Koschella, J. Schaller, photometric determination of polymer-supported F. Meister, Solvent-free synthesis of 6-deoxy-6- amino groups. Analyst 114, 1147-1150 (1989). (ω-aminoalkyl) amino cellulose. J. Appl. Polym. [14] S. S. Chu, S. H. Reich, NPIT: a new reagent for Sci. 133, DOI: 10.1002/APP.43987 (2016). quantitatively monitoring reactions of amines in [24] L. Jin, W. Li, Q. Xu, Q. Sun, Amino-functionalized combinatorial synthesis. Bioorg.Med. Chem. nanocrystalline cellulose as an adsorbent for an- Lett. 5, 1053-1058 (1995). ionic dyes. Cellulose 22, 2443-2456 (2015).

120 LENZINGER BERICHTE 94 (2018) 115 – 122

Annex

Table A1: Structure, classification, and solvents for the dyes applied.

Dye / Classification Dissolved in Structure

70% Water Coomassie-Brilliant-Blue G250/ 20% Methanol Triphenylmethane 10% Acetic acid

Eriochrome black T/ Water Azo

Congo red/ 50% Water Azo 50% Ethanol

Crystal violet/ 70% Water Triphenylmethane 3% Methanol

Methylene blue/ 70% Water Indamin 30% Methanol

Thymol blue/ 70% Water Triphenylmethane 30% Methanol

Methyl red/ Ethanol Azo

Ruhemans purple/ 50% Water Ninhydrin 50% Ethanol

Fluorescence Measurements In Figure A1 the slopes for the quantum yield cal- Fluorescence measurements were conducted in water- culations are plotted for the standards and the labeled citric acid buffer and a diluted aqueous hydrochloric amino cellulose investigated. The quantum yields can acid solution at pH =3.1. Due to self-quenching ini- be calculated using equation A1: tiated by a photoinduced electron transfer the fluorescence dye employed needed to be kept at a low pH value [1]. When the amino anthracene nitrogen becomes protonated, this effect vanishes and the Here the subscripts ST and X corresponds to the stan- fluorescence of the compound can be measured. dard and sample respectively. φ is the fluorescence

121 LENZINGER BERICHTE 94 (2018) 115 – 122

quantum yield and Grad is the gradient from the inte- The average fluorescence quantum efficiency of the grated fluorescence intensities against the absorbance labeled amino cellulose in aqueous citric acid and di- as seen in the plot (Figure A1), η corresponds to the luted hydrochloric acid at pH = 3.1 is φAmcell = 45 %. refractive indices of the solvents used for the measu- Knowing these emission properties, it was possible to rements [2]. The quantum yields for anthracene and directly measure the amount of labeled amino cellulose 9,10-di-phenylanthracene are given with φanthr. = 0.27 that was left over in the solution after the absorption and φ9,10-diph = 0.95 [3,4]. experiments with different pulp samples (see Table 1).

anthracene EtOH 9,10 diphenylanthracene EtOH labeled aminocellulose water/HCI labeled aminocellulose water/citric acid

Equation y = a + b*x Adj. R-Square 1 0.972 0.999 1 Value st. error anthracene in EtOH Intercept 891.174 184.239 anthracene in EtOH Slope 7.45E6 64514.44 9.10 dipehnylanthracene in EtOH Intercept 5159.01 2175.698 9.10 dipehnylanthracene in EtOH Slope 2.836E7 2.392E6 labeled aminocellulose in water/HCI Intercept 735.42 452.625 pH=3.1 labeled aminocellulose in water/HCI Slope 1.352E7 194313.074 pH=3.1

integr. fluorescence area in a.u. integr. labeled aminocellulose in water/ Intercept 329.07 105.722 citric acid pH=3.1 labeled aminocellulose in water/ Slope 1.334E7 43434.568 citric acid pH=3.1

absorption ideal. 350 nm

Figure A1: Integrated fluorescence intensities versus the idealized absorptions at 350 nm of two standards (anthracene and 9,10-diphenylanthracene in ethanol, EtOH) and the fluorescence dye labeled 6-deoxy-6-(2-aminoethyl)amino cellulose in water- citric acid and diluted hydrochloric acid solution at pH = 3.1. The resulting gradients can be used in equation A1 to calculate the quantum yields.

References

[1] R. A. Bissell, E. Calle, A. Prasanna de Silva, [2] A. T. R. Williams, S. A. Winfield, J. N. Miller, S. A. de Silva, H. Q. Nimal Gunaratne, J.-L. Habib- Analyst, 108, 1067-1071 (1983). Jiwan, S. L. Annesley Peiris, R. A. D. Dayasiri [3] W. R. Dawson, M. W. Windsor, J. Phys. Chem. Rupasinghe, T. K. Shantha, D. Samarasinghe, 72, 3251-3260 (1968). K. R. A. Samankumara Sandanayake, J.-P. Sou- [4] J. V. Morris, M. A. Mahaney, J. R. Huber, J. million, J. Chem. Soc. Perkin Trans. 2, 1559- Phys. Chem. 80, 969-974 (1976). 1564 (1992).

122 LENZINGER BERICHTE 94 (2018) 123 – 130

Possibilities for Raw Material Base Expansion for Lyocell Applications by Enzymatic Pulp Treatment

Birgit Kosan1,* and Frank Meister1

1 Thüringisches Institut für Textil- und Kunststoff-Forschung e.V., Breitscheidstr. 97 D-07407 Rudolstadt, Germany * Contact: (+49) 3672 379 220, Fax: (+49) 3672 379 379, e-mail: [email protected]

Abstract

The prospects of an enzymatic modification procedure were studied for adjustment of alternative pulp properties in Lyocell fibres spinning process. Ultimately, an enzymatic treatment was developed which is adaptable to the Lyocell process technology using cellulases with well-defined endo- and exo-activities for an intended reducing of the cellulose DP as well as for an improvement of the pulp solubility in NMMO. The DP reduction of different tested pulp qualities (paper, viscose and annual plant pulps) was attained between 20 and 25%, with regard to the starting DP of the used pulp sample. Thereby, very small pulp loss was detected. Because of the significant improvement of the pulp solubility by the enzymatic modification, spinning dopes with very good solution states could be received, even when paper grade pulps with lower contents of α-cellulose were applied. Exemplary for a softwood TCF paper grade pulp, tests for first upscaling in staple fibre and filament preparation could be successfully carried out with integration of the developed enzymatic modification step. The prepared fibre and filament samples showed good textile physical properties. The developed, innovative enzymatic modification step has a great potential for further expansion of the raw material base in Lyocell process. This could cause to cost savings by applicability of less expensive cellulose raw materials, but it provides also broader possibilities for usage of alternative cellulose sources from recycling materials or annual plants, outside of cotton.

Keywords: Enzymatic pulp treatment, Lyocell process, pulp base expansion, DP adaption, solution state

Introduction

Because of the predicted increase of the world textile at cotton cultivation [2]. The resulting demand is re- fibre consumption at simultaneous stagnation of cotton ferred to as “cellulose gap”. It opens up the possibili- cultivation, the demand for cellulose textile fibres will ty, but also the necessity for a significant improve- continue to grow, and it is expected an annual increase ment of the capacities for production of sustainable of the production of textile fibres of about 3.1% up to cellulosic regenerated fibres [1,3]. the year 2030 [1]. This predicted, disproportionately The Lyocell process using direct dissolution of cellu- high increase of the fibre demand by around 2.8 Mio t/a lose in N-methylmorpholine-N-oxide monohydrate is neither feasible by appropriate improvement of the (NMMO) has a special potential for sustainable cellu- production of synthetic fibres due to their different lose fibre processing [4]. The solvent is non-toxic and moisture management nor by a further increase of the nearly completely recyclable [5]. However, the Lyocell cotton production, because of limitation or reduction fibre processing requires the use of special dissolving of the cultivated amounts in favour of also increasing pulp qualities which were adapted directly for the pro- world food demand as well as comparatively unfa- cess. It would be of great interest to extend the range vourable farming criteria, i.e. high water consumption of cellulosic raw materials for Lyocell fibre spinning or usage of genetically modified seeds and pesticides applications. The usage of low-cost pulp qualities, i.e.

123 LENZINGER BERICHTE 94 (2018) 123 – 130

paper grade pulps, cellulose from annual plants [6] or Preparation of cellulose dopes fibre raw material from recycled textile sources could be The preparation of cellulose dopes was carried out very interesting on grounds of cost but also from the using a special vertical kneader system, linked with a point of view of a circular flow economy. The paper RHEOCORD 9000 (HAAKE). Temperature, torque will report and discuss possibilities of an enzymatic moment and revolutions per minute (rpm) vs. time paper pulp treatment, adaptable in the Lyocell process, were recorded on-line. The dopes were prepared, start- using suitable cellulases with special ratios of exo- ing from an aqueous suspension of the treated pulp activities for improvement of the accessibility of the cel- in 50% aqueous NMMO, by removal of the excess lulose molecules [7] and endo-activities for targeted water at elevated temperatures, higher shearing stress reduction of their long-chain cellulose parts. and low pressure during the dissolution processes (80-95 °C mass temperature, 800-40 mbar pressure, Materials and Methods 5-20 rpm). 0.5% (w/w) propylgallate, with regard to cellulose, were used for stabilisation of the NMMO The investigations were carried out using softwood solutions. After the finishing of water removal (achiev- TCF paper grade pulp, (NH-P). The determined ana- ing a NMMO monohydrate state), an after-dissolution lytical data of the pulp are listed in Table 1. step (60 min, 90 °C mass temperature, 250 mbar) followed for homogenisation of the prepared dope. Table 1: Analytical data of the used paper grade pulp sample A first upscaling in 4 kg dope scale was carried out using planetary mixing machine PML 40 (Netzsch- Pulp sample NH-P Cuoxam-DP 763 Feinmahltechnik GmbH). α-cellulose content [%] 85.5 Content of carboxyl groups [μmol/g] 33.2 Spinning trials Content of carbonyl groups [μmol/g] 27.6 First spinning tests were carried out by dry-wet spinning experiments for preparation of staple fibres of The used enzyme samples were cellulase test samples about 1.7 dtex fineness using a laboratory piston spin- with selected endo and combined exo activities, which ning equipment, which is described in former publi- were kindly selected and delivered by Biopract cation. [8] Spinning nozzles, containing 30 holes with GmbH. capillary diameters of 100 μm were used for all labora- N-Methylmorpholine-N-oxide (NMMO) was used as tory spinning experiments. The spinning temperatures 50% aqueous solution, manufactured and supplied were selected in each case according to the rheologi- by BASF, without any further pre-treatment. cal properties of the used cellulose dopes. All other used chemicals were obtained in analytical Further semi-technical spinning tests using spinnerets grade, respectively and were used without any further with 6 x 80 capillaries (90 μm outlet diameter) were pre-treatment. carried out for investigation of the spinning behaviour and stability. These trials were used for preparation of Enzymatic treatment staple fibres and multifilament samples. The enzymatic treatment of the pulp samples was separately carried out before the pulp was admixed to Cellulose characterisation the aqueous NMMO solution and could be smoothly The virgin and enzymatic treated pulp samples as well integrated into the existing NMMO technology. as the cellulose samples regenerated from the dopes The treatment procedure took place in aqueous medium and from the spinning tests dissolved in Cuoxam were using the listed enzyme concentrations (see Table 2), characterised by capillary viscosity for determination of referred to cellulose, at pH of 5.5 during 60 minutes the average degree of polymerisation (Cuoxam-DP). treatment time at 50 °C. The treatment was stopped by The α-cellulose content was determined by investiga- increase of the pH up to 9. Following, the treated pulp tion of the pulp amounts which are resistant to 17.5% was pressed out from the excess water down to about caustic soda at 20 °C. 30% solid content and introduced in aqueous NMMO The determination of carboxyl group contents has for further processing. been carried out through complexometric titration of For analytical characterisation, the treated pulp samp- zinc ions after removing of the metal ions from the les were washed by deionised water up to neutral pH, cellulose at first and adding of zinc acetate solution in pressed out from the excess water and air dried. a second step. The carbonyl group contents were analysed by measurement of the absorbance at 530 nm after reaction with 2, 3, 5-triphenyltetrazoliumchloride solution.

124 LENZINGER BERICHTE 94 (2018) 123 – 130

The details of this cellulose characterisation were de- Results and Discussion scribed in former publications. [9,10] The possibilities of the enzymatic treatment should be Analytical characterisation of exemplary discussed in this paper for the softwood TCF cellulose solutions paper grade pulp sample (NH-P). The untreated pulp Optical characterisation of cellulose solutions was car- sample was not suitable for direct use in Lyocell pro- ried out by means of polarisation microscopy (ZEISS cessing because of the comparably high Cuoxam-DP Axiolab). Furthermore measurements by laser diffrac- of 763 and the unfavourable dissolution behaviour in tion (HELOS particle size equipment, SYMPATEC) NMMO. were used for determination of the particle distribution The enzymatic treatment of the pulp sample with cellu- and particle content of prepared spinning dopes [11]. lase samples P1, P4, P10 and P11 containing different The determination of the solids contents in the solu- mixtures of enzymatic activities resulted in a reduction tions was performed by means of weighing precipitated of the Cuoxam-DP of the pulp sample between 19 and films, after exhaustive solvent extraction and drying. 23%. The attained DP values of the modified pulp The rheological measurements were realised as de- samples were between 589 and 619 in an appropriate scribed in the following. area for the preparation of Lyocell fibres. The content of carbonyl groups of the pulp sample increases, de- Rheological characterisation pending on the DP reduction and the used enzyme The rheological characterisation of cellulose solutions type and applied concentration (Table 2). This increase was performed using a rheometer HAAKE MARS of the carbonyl group contents should indicate the ex- with cone / plate measuring system (4° angle geometry) tent of the chain scission and should be significantly and electrically heated cone & plate unit with active influenced by the resulting molecular weight distribu- cone heater. tion especially on the low molecular amounts. Zero shear viscosities were calculated from creep attempts in the rotation mode at shear stress of 90 Pa. Table 2: Analytical data of pulp sample NH-P after enzyma- The determined cellulose solutions showed signifi- tic treatment cant viscoelastic behaviour. A characterisation of the Content of viscoelastic properties was realised in the oscillation carbonyl Cuoxam Reduction groups mode. Oscillation tests were carried out as frequency Sample / Enzym dosage* -DP of DP [μmol/g] sweeps (0.046-14.7 Hz, deformation: 0.07) at different NH-P (untreated) 763 27.6 temperatures (60 / 85 / 110°C). WLF-transformation NH-P / 0,3% P4 619 18.9 61.4 was used for calculation of master curves and relaxa- NH-P / 1% P10 612 19.8 51.3 tion spectra at 85°C reference temperature. This method NH-P / 1% P11 590 22.7 48.6 permits an interpolation of the frequency / angular NH-P / 0,5% P1 + 0,3% P4 589 22.8 88.2 * Enzyme concentration referred to pulp amount rate range over several decades and is an evaluated method for the determination of viscoelastic properties of polymer solutions. Details of the rheological Dissolution and shaping tests dope characterisation had been described in former in laboratory scale publications. [9,12] The enzymatic modified pulp samples could be used for dope preparation in laboratory scale with cellulose Fibre characterisation concentrations between 12 and 12.5% (w/w). The micro- The textile-physical fibre parameters were determined scopic images of the prepared dopes showed a strongly according to the following methods: improved solution states without any undissolved fibre residuals. Contrary to this, the dissolution of the un- • fineness according to DIN EN ISO 1973, modified pulp sample resulted in a very bad solution • tenacity and elongation at break according to DIN state, even at reduced cellulose concentration of 9% EN ISO 5079 and (w/w), containing a lot of undissolved fibre fragments. • loop tenacity according to DIN 53843, part 2. Figure 1 shows the microscopic images of a 9% dope from the unmodified pulp sample compared to a 12% The filament samples were tested according to: dope using enzymatic (P11) modified pulp sample. Be- • yarn fineness according to DIN EN ISO 2060 and cause of the high particle content with a large number - maximum tensile strength and elongation according of particles with high aspect ratio, a particle character- to DIN EN ISO 2062. isation of the 9% dope with laser diffraction was not possible. In contrast, very small particle contents (< 20 ppm)

125 LENZINGER BERICHTE 94 (2018) 123 – 130

are pictured in Table 3 and Figures 1 and 2. The spinning dope which was prepared using the unmodified pulp sample showed lower viscosity and moduli curves, because of the lower cellulose concentration and the incomplete dissolution state. The dopes were tested regarding their formability to staple fibres using laboratory piston spinning equip- 9% unmodified pulp 12% NH-P (1% P11 ment (capillary diameter: 100 µm, 30 capillaries). The NH-P in NMMO modified) in NMMO applied spinning conditions and achieved fibre proper- Figure 1: Microscopic images of prepared spinning dopes in ties are listed in Table 4. NMMO (transmitted light, polarization microscopy, Spinning dope 1, prepared by using the unmodified 10x-objective) pulp sample, showed a very unstable spinning behaviour. Alongside many capillary breaks, the spinning with low particle sizes (up to 51 μm in maximum) were pressure increased from 32 to 38 bar already during detected by laser diffraction at the prepared dopes from the short time of the laboratory spinning test of about enzymatic modified pulp samples. 30 minutes. This pressure increase should be caused The rheological properties of the prepared spinning by accumulation of particles on the safety filter device dopes, applying the enzymatic modified pulp samples, because of the worse dope quality containing undis- are in a well-suitable area for Lyocell process tech- solved fibre residuals or rather gel contents. nology. The results of the rheological characterization

Table 3: Results from analytical characterisation of cellulose dopes using pulp sample NH-P

Sample 1 2 3 4 5 Pulp treatment without 0.3% P4 0.5% P1 / 0.3% P4 1% P10 1% P11 Cuoxam-DP of applied pulp sample 763 619 589 612 590 Cellulose concentration % 9.1 11.8 12.5 12.2 12.0 Zero shear viscosity (85 °C) Pas 7,295 12,130 14,160 11,570 9,587 Angular velocity (cross over) 1 rad/s 1.27 1.37 1.34 1.68 1.88 Storage modulus (cross over) 1 Pa 1,193 2,126 2,339 2,360 2,223 Plateau modulus 1 Pa 16,100 24,500 26,700 25,900 23,400 Rheological polydispersity 4,5 4.5 4.8 4.8 4.7 Relaxation time λm at H*m 2 s >100 41 87 19 20 1 calculated from master curves (Figure 2) 2 calculated from weighted relaxation time spectra (Figure 3)

Table 4: Spinning conditions and achieved fibre properties using pulp sample NH-P

Sample 1 2 3 4 5 Pulp treatment without 0.3% P4 0.5% P1 / 0.3% P4 1% P10 1% P11 Cuoxam-DP of applied pulp sample 763 619 589 612 590 Spinning conditions: Spinning temperature °C 87 89 91 89 87 Spinning pressure bar 32-38 29 32 33 23 Spinning speed m/min 30 30 30 30 30 Spinning behaviour unstable stable stable stable stable Fibre testing: Fineness dtex 1.81 1.75 1.83 1.79 1.40 Tenacity, cond. cN/tex 30.8 35.4 32.7 34.2 33.8 Elongation, cond. % 12.2 14.4 14.9 13.3 13.5 Loop tenacity cN/tex 14.7 14.7 15.3 17.3 16.0 E-modulus 0.5-0.7% cN/tex 795 922 779 882 855 Cuoxam-DP fibre 703 567 572 560 545 Water retention value (WRV) % 95.6 87.4 89.7 91.7 82.6

126 LENZINGER BERICHTE 94 (2018) 123 – 130

Figure 2: Master curves of spinning dopes using enzymatic treated and unmodified pulp sample NH-P Triangles: Storage moduli Squares: Loss moduli Open circles: Complex viscosities

Figure 3: Weighted relaxation time spectra of spinning dopes using enzymatic treated and unmodified pulp sample NH-P

The spinning tests using the dopes from enzymatic slightly below the level of Lyocell fibres. It illustrates, modified pulp samples showed a very stable spinning that the prepared fibres from paper grade pulps showed behaviour. The Cuoxam-DP values of the prepared fibres slightly increased WRV compared with Lyocell fibres, were comparable between 545 and 572. The textile prepared using typical Lyocell pulps. [13] physical properties of the prepared fibres were only

127 LENZINGER BERICHTE 94 (2018) 123 – 130

First upscaling tests parable to typical Lyocell pulps. Semi-technical spin- First upscaling tests for the enzymatic treatment (1% ning tests (6 x 80 capillaries, 90 μm outlet diameter) P11) of the pulp sample NH-P and following dope for investigation of the spinning behaviour and stability preparation were carried out for staple fibre and fila- as well as the preparation of staple fibres and multi- ment preparation. Both the enzymatic treatment and the filament yarn were carried out using 12% (w/w) cellu- dope preparation were carried out firstly with discon- lose dopes. The rheological properties of the spinning tinuous batch trials. The preparation of the spinning dopes prepared with these batch trials were comparable dopes was possible using dissolution conditions com- to the dopes from the lab tests (Figures 4 and 5).

Figure 4: Master curves of spinning dopes comparison of dopes from lab scale and larger batch Triangles: Storage moduli Squares: Loss moduli Open circles: Complex viscosities

Figure 5: Weighted relaxation time spectra – comparison of dopes from lab scale and larger batch

128 LENZINGER BERICHTE 94 (2018) 123 – 130

Tests for staple fibre preparation as well as filament of the prepared staple fibres as well as the determined yarn preparation were carried out at spinning speeds of single fibre and yarn values of the prepared multifila- 33 m/min., spinning pressures between 53 and 57 bar, ments. The detected fibre properties confirm the high spinning temperatures of 85°C, at typical nozzle ser- potential of the tested enzymatic modification step to vice lifetimes without significant improved pressure adjust alternative cellulose resources for Lyocell appli- increase. cations. The prepared fibres and filaments revealed good textile physical properties. Table 5 illustrates the properties

Table 5: Fibre and filament properties from batch trials using pulp sample NH-P (P11 modified)

Fibre properties Fibre properties Yarn properties staple fibres filaments filaments Fineness dtex 1.46 1.46 641 Tenacity, cond. cN/tex 38.6 41.3 40.2 Elongation, cond. % 11.8 7.6 4.7 Loop tenacity cN/tex 13.6 9.5 28.4 E-modulus 0.5-0.7% cN/tex 963 1,317 1,930 Cuoxam-DP fibre 513

Fibre and filament material from batch tests

Conclusions process technology. It could cause for cost savings when less expensive cellulose raw materials (paper An enzymatic modification using cellulases with ad- pulps) are applied, but it may provide also broader apted exo- and endo-activities is a suitable approach options for use of pulps extract from recycled cellulo- for adjustment of the required average degree of poly- se fabrics or from annual plant fibres next to cotton. merisation and for improved solubility properties of alternative pulp resources for Lyocell fibre spinning. Acknowledgements The tested modification step should be adaptable in the Lyocell process. The DP reduction of different tested These works were financially supported by the German pulp qualities was detected between 20 and 25%, with Federal Ministry of Economics and Energy (BMWi regard to the DP of the non-modified pulp sample. Be- No. MF140191). The authors thank for the granted side the DP reduction, the tested enzyme samples with support. adjusted exo- and endo-activity surprisingly effected Furthermore the authors would like to thank Dr. Eng. also a significant improvement of the pulp solubility. M. Gerhardt and V. Pelenc (Biopract GmbH) for the Therefore, spinning dopes with very good dissolution assistance and providing of the enzyme samples. behaviour were received, even when paper grade pulps with lower contents of α-cellulose are used. References Successful tests for first upscaling for staple fibre and filament preparation could be carried out during appli- [1] F. M. Hämmerle, The Cellulose Gap (The future cation of softwood TCF paper grade pulp (Cuoxam- of cellulose fibres), Lenzinger Berichte, 89, DP: 763, content of α-cellulose: 85.5%), and integra- 2011, 12-21 tion of the developed enzymatic modification step. The [2] Gherzi, The future challenges for cotton avail- prepared fibre and filament samples showed good textile ability: Chance for Man-made cellulose fibers?, physical properties for typical Lyocell fibrous regen- Allgemeiner Vliesstoff-Report 2 / 2011 erates (tenacity in conditioned state: 39-41 cN/tex). [3] D. Eichinger, A Vision of the World of Cellulo- The use of such an innovative modification step can sic Fibers in 2020, Lenzinger Berichte, 90, 2012, significantly extend the raw material base for Lyocell 1-7

129 LENZINGER BERICHTE 94 (2018) 123 – 130

[4] H. Harms, Potentiale einer neuen Fasergenera- solution and processing properties of cellulose in tion, Lenzinger Berichte, 81, 2002, 27-34 ionic liquids, Lenzinger Berichte, 90 (2012) 76-84 [5] W. Kalt, B. Zauner, Das Lenzing Lyocell Projekt – [10] F. Meister, B. Kosan, A tool box for characteriza- Start in ein neues Zeitalter der großtechnischen tion of pulps and cellulose dopes in Lyocell tech- Zellulosefaserherstellung in Europa, Lenzinger nology, Nordic Pulp &Paper Research Journal Berichte, 80, 2001, 10-12 30 (1), 2015, 112-120 [6] J. Paulitz, I. Sigmund, B. Kosan, F. Meister, Lyo- [11] B. Kosan, Ch. Michels, Particle analysis by laser cell fibers for textile processing derived from or- diffraction – application and restrictions in the ganically grown hemp, Procedia Engineering, Lyocell process, Chemical Fibers International, 200 (2017) 260-268 49, 1999, 50-54 [7] F. Meister, Ch. Michels, H. Kramer, Regenerated [12] C. Michels, B. Kosan, Contribution to dissolution Cellulose Moulding and Process for Producing it, state of cellulose in aqueous amine oxide charac- WO 1996014451, 1996 terized by optical and rheological methods, Len- [8] B. Kosan, C. Michels, F. Meister, Dissolution and zinger Berichte, 82, 2003 128-135 forming of cellulose with ionic liquids, Cellulose, [13] Ch. Michels, B. Kosan, Beitrag zur Struktur von 15 (2008) 59-66 Lyocellfasern, ersponnen aus Aminoxidhydraten [9] B. Kosan, K. Schwikal, F. Meister, Effects of pre- bzw. ionischen Flüssigkeiten, Lenzinger Berichte, treatment and dissolution conditions for improved 86, 2006, 144-153

130